Time Ch1 Ch2 Ch3 Ch4 Ch5

Monday, September 21

06:00 pm-06:30 pm When STEM and Management mingle harmoniously: a conversation with Dr. Elena Grifoni Winters

Tuesday, September 22

09:00 am-10:00 am PT1: Plenary Talk1 - Peter Knott (FHR, Germany): Radar Research at Fraunhofer FHR - Challenges and Way Ahead
10:00 am-10:20 am Coffee Break
10:20 am-12:00 pm TM-L1: Learning techniques for radar and EW TM-L2: Passive Radars TM-L3: Phased Array Antennas TM-L4: Radar Detection TM-SS1: Urban remote sensing using SAR.
12:00 pm-02:15 pm TM-P1: PSS1-Resource management for cognitive radar systems,
TM-P2: Multistatic and passive radars,
TM-P3: Wideband antennas and phased array,
TM-P4: Cognitive radars,
TM-P5: Target detection and clutter suppression,
TM-P6: Space-based/Airborne radars
02:15 pm-03:00 pm Opening ceremony
03:00 pm-04:00 pm Inaugural Plenary Talk - Marco De Fazio (Leonardo Spa, Italy): Radars at Leonardo Company: a 70 years long heritage in land, naval, airborne, and spaceborne systems. Looking backward, moving forward.
04:00 pm-04:20 pm Coffee Break
04:20 pm-06:00 pm TA-SS2: Topics, trends and challenges in cognitive radars TA-L2: Multistatic radars TA-SS3: Quantum radar: real world experiments and new theory TA-L4: Short range civilian radars TA-SS4: Distributed SAR systems and missions

Wednesday, September 23

09:00 am-10:40 am WM-SS5: Sparse array design techniques for radar applications WM-L2: Radar waveform design/optimization WM-SS6: Emerging technologies in automotive radars WM-L4: Classification and recognition of drones WM-SS7: Multitemporal SAR image processing and analysis
10:40 am-11:00 am Coffee Break
11:00 am-12:40 pm WM-L6: Compressed sensing and sparsity exploitation WM-SS8: Coexistence of radar and communication systems WM-SS9: Interference in automotive radars WM-SS10: Detection, tracking, and classification of small drones WM-L10: Space-based radar systems and missions
12:40 pm-02:00 pm Lunch Break
02:00 pm-03:00 pm Plenary Talk 2 - Markus Gardill (University of Würzburg. Germany): State-of-the-Art Automotive Radar System Architectures - and What Else We Can Do with Them.
03:00 pm-04:40 pm WA-SS11: Emerging technologies for radar applications: compressive sensing meets machine learning WA-SS12: Dual-function radar/communication systems WA-L3: Automotive radars WA-L4: Radar for structural monitoring WA-L5: Space-based/Airborne radar remote sensing applications
04:40 pm-05:00 pm Coffee Break
05:00 pm-05:20 pm WA-D1: DEMO 1 WA-D2: DEMO 2 WA-D3: DEMO 3    
05:20 pm-07:20 pm WA-P1: PSS2 - Advanced radar waveform design strategies,
WA-P2: Localization and estimation,
WA-P3: Waveform diversity and MIMO radars,
WA-P4: Automotive radar technology and signal processing,
WA-P5: Experimental short-range target detection and classification,
WA-P6: Radar tracking and track-before-detect

Thursday, September 24

09:00 am-10:40 am ThM-L1: DOA Estimation ThM-L2: Joint radar and communications ThM-L3: Radar technology ThM-L4: Radar phenomenology and modeling ThM-SS13: SAR meets AI
10:40 am-11:00 am Coffee Break
11:00 am-12:40 pm ThM-SS14: Multisensor multitarget tracking in surveillance applications ThM-SS15: Multi-function spectral system co-design ThM-L8: Radar classification ThM-SS16: Ground based radar remote sensing of clouds and precipitation ThM-L10: High resolution radar imaging
12:40 pm-02:00 pm WIE Panel: Gender bias in the workplace
02:00 pm-03:00 pm Plenary Talk 3 - Mahta Moghaddam (University of Southern California, USA): Microwave Sensing for Medical Imaging and Monitoring of Thermal Therapies: Accelerated Inverse Scattering via Learning.
03:00 pm-04:40 pm ThA-L1: Radar tracking ThA-L2: Software-defined, small radar architectures, prototypes ThA-SS17: Radar for health monitoring and biomedical applications ThA-SS18: Satellite sensing of the atmosphere: radar technologies and methods for advancing atmospheric and climate science. ThA-SS19: Millimeter-wave synthetic aperture radar
04:40 pm-05:00 pm Coffee Break
05:00 pm-05:20 pm ThA-D1: DEMO 1 ThA-D2: DEMO 2 ThA-D3: DEMO 3    
05:20 pm-06:40 pm ThA-P1: PSS3 - Multitarget tracking,
ThA-P2: PSS4 - Multisensor multitarget tracking,
ThA-P3: PSS5 - Radar networks for climate change,
ThA-P4: Indoor, GPR and through-the-wall radars,
ThA-P5: ECCM, defense and technology,
ThA-P6: High resolution SAR/ISAR,
ThA-P7: Detection and estimation

Monday, September 21

Monday, September 21 6:00 - 6:30 (Europe/Rome)

When STEM and Management mingle harmoniously: a conversation with Dr. Elena Grifoni Winters

Chair: Alfonso Farina (Leonardo Company Consultant, Italy)

Tuesday, September 22

Tuesday, September 22 9:00 - 10:00 (Europe/Rome)

PT1: Plenary Talk1 - Peter Knott (FHR, Germany): Radar Research at Fraunhofer FHR - Challenges and Way Ahead

Chair: Pierfrancesco Lombardo (University of Rome La Sapienza, Italy)

Tuesday, September 22 10:00 - 10:20 (Europe/Rome)

Coffee Break

Tuesday, September 22 10:20 - 12:00 (Europe/Rome)

TM-L1: Learning techniques for radar and EW

Room: Ch1
Chairs: Alexander Charlish (Fraunhofer FKIE, Germany), Visa Koivunen (Aalto University, Finland)
10:20 Reinforcement Learning-Based Joint Adaptive Frequency Hopping and Pulse-Width Allocation for Radar anti-Jamming
Ailiya Ailiya, Wei Yi and Ye Yuan (University of Electronic Science and Technology of China, China)
It is shown that frequency hopping and pulse-width allocation strategy can provide enhanced anti-jamming performance for the radar systems. The current anti-jamming methods often have difficulty in adapting their policy to the complicated and unpredictable jamming environment. To address this limitation, a reinforcement learning-based joint adaptive frequency hopping and pulse-width allocation scheme is proposed. By applying the reinforcement learning, the radar can learn the optimized anti-jamming policy by interacting with the environment and requires little prior information. In the proposed scheme, we first establish a reward model to quantify the performance of radar anti-jamming decisions. Then, the radar anti-jamming decision process is modeled as a Markov decision process. As one of the widely-used reinforcement learning algorithms, the Q-learning, which can converge to the optimized policy with probability 1, is utilized to learn the optimized radar anti-jamming policy in the context of lacking a perfect environmental knowledge. Numerical results are shown to verify the effectiveness of our proposed strategy.
10:40 CNN-LSTM Based Approach for Parameter Estimation of K-clutter plus Noise
Taha Hocine Kerbaa and Amar Mezache (University of Mohamed Boudiaf, M'sila. Algeria); Fulvio Gini and Maria S. Greco (University of Pisa, Italy)
This paper concerns the problem of estimating the parameters of the K plus noise distribution. In a previous work, it has been shown that, in the multilook scenario, the modified fractional order moment estimator (MFOME) has about the same estimation accuracy as the [zlog(z)] method, but lower computational complexity. However, significant estimation errors have been observed in the single look scenario, low sample size, and large values of the K-distribution shape parameter. Moreover, the computational complexity of the [zlog(z)] estimator discourages its implementation in practical applications. The aim of this work is to estimate the shape parameter of the K-distribution with reduced computational complexity. The problem can be formulated as a supervised many-to-one sequence prediction. We propose here a hybrid model including convolutional and long-short-term-memory (LSTM) neural networks (NN). Estimation performance is investigated by processing both simulated and real clutter data.
11:00 Deep Learning for Accurate Indoor Human Tracking with a mm-Wave Radar
Jacopo Pegoraro, Domenico Solimini, Federico Matteo, Enver Bashirov, Francesca Meneghello and Michele Rossi (University of Padova, Italy)
We address the use of backscattered mm-wave radio signals to track humans as they move within indoor environments. The common approach in the literature leverages the extended Kalman filter (EKF) method, which however undergoes a severe performance degradation when the system evolution model is highly non-linear or presents long-term time dependencies among the system states. In this work, we propose an original model-free tracking procedure based on denoising autoencoders and sequence-to-sequence neural networks, showing its superior performance with respect to state-of-the-art methods. Our architecture can be trained in either a supervised or unsupervised manner, trading tracking accuracy for flexibility. The proposed system is tested on our own measurements, obtained with a 77 GHz radar on single and multiple subjects simultaneously moving in an indoor space. The results are compared against the ground truth trajectories from a motion tracking system, obtaining average tracking errors as low as 12 cm.
11:20 Discrimination of Air Breathing Targets and Ballistic Missiles Using Deep Learning
Alfonso Farina (Leonardo Company Consultant, Italy); Massimo Loffreda (Leonardo spa, Italy); Luca Timmoneri (Leonardo Spa, Italy)
Cognitive radar differs from traditional radar as well as from active phased array radar because of their capability in developing rules of behaviour in a self-organized manner. This is obtained by the so-called learning from experience process that results from continue interactions with the environment after a huge training phase on synthetic data. In this paper we present the results achieved applying Data Learning techniques to one of the most complex function for a surveillance radar: the classification and identification of air target with particular attention to the discrimination between air breathing targets (ABTs) and ballistic missiles (BMs).
11:40 Deep Learning for Radar Signal Detection in Electronic Warfare Systems
Mustafa Nuhoglu (Istanbul Technical University & ASELSAN, Turkey); Yasar Kemal Alp (ASELSAN Inc., Turkey); Fatih Cagatay Akyon (Bilkent University, Turkey)
Detection of radar signals is the initial step for passive systems. Since these systems do not have prior information about received signal, application of matched filter and general likelihood ratio tests are infeasible. In this paper, we propose a new method for detecting received pulses automatically with no restriction of having intentional modulation or pulse on pulse situation. Our method utilizes a cognitive detector incorporating bidirectional long-short term memory based deep denoising autoencoders. Moreover, a novel loss function for detection is developed. Performance of the proposed method is compared to two well known detectors, namely: energy detector and time-frequency domain detector. Qualitative experiments show that the proposed method is able to detect presence of a signal with low probability of false alarm and it outperforms the other methods in all signal-to-noise ratio cases.

TM-L2: Passive Radars

Room: Ch2
Chairs: Fabiola Colone (Sapienza University of Rome, Italy), Hugh Griffiths (University College London, United Kingdom (Great Britain))
10:20 Detecting and Tracking a Small UAV in GSM Passive Radar Using Track-before-Detect
Benjamin Knoedler (Fraunhofer FKIE, Germany); Christian Steffes (Fraunhofer Institute for Communication, Information Processing and Ergonomics FKIE, Germany); Wolfgang Koch (Fraunhofer FKIE & University of Bonn, Germany)
The detection and tracking of consumer grade unmanned aerial vehicles is of great interest in recent years due to threats arising from the widespread availability of such devices. Passive radar is considered as a sensor system capable of contributing to this task. Weak target reflections together with disadvantages of traditional processing schemes make detecting and tracking small drones with passive radar a challenge. The approach of Track-before-Detect applies unthresholded measurements to overcome these actualities and possibly detect previously undetectable targets. In this work it is shown through evaluation of experimental data that Track-before-Detect methods are very capable to successfully detect and track small drones in GSM passive radar.
10:40 Localization of Micro Unmanned Aerial Vehicles Using Digital Audio Broadcast Signals
Samuel Welschen (ETH Zürich, Switzerland); Christof Schüpbach (Armasuisse Science+Technology, Switzerland); Stephen Paine (Armasuisse Science+Technology, Switzerland & University of Cape Town, South Africa); Urs Böniger (Armasuisse Science+Technology, Switzerland); Pascal Leuchtmann (ETH Zürich, Switzerland); Juerg Leuthold (ETH Zurich, Switzerland)
Localizing micro unmanned aerial vehicles (MUAV) using passive radar systems makes it possible to counter the threat they pose to critical infrastructure. In this paper, we discuss an efficient signal processing chain for their localization and demonstrate the detection and localization of cooperative MUAVs using a five-element array and Digital Audio Broadcast (DAB) signals. The MUAVs were detected up to their maximum flight distance from the receiver of 2.6 kilometers. By estimating the direction of arrival of the target echoes in addition to the bistatic range and Doppler frequency shift, the MUAVs could be localized with a mean position error below 90 meters.
11:00 Polarimetric Antenna Diversity for Improved Reference Signal Estimation for Airborne Passive Radar
Philipp M. Wojaczek, Diego Cristallini, Jochen Schell and Daniel O Hagan (Fraunhofer FHR, Germany); Ashley Summers (Defence Science and Technology Group, Australia)
Passive radar utilizes signals from broadcast, communication, or other transmitters of opportunity for the applications of target detection and imaging. For these applications knowledge of the transmitted signal for the correlation with the surveillance signal is a crucial factor. If digital signals are exploited for passive radar, it is possible to estimate and restore the transmitted signal. The error rate in the estimated transmitted signal is inversely related to the signal-to-noise ratio of the direct signal. The greater the amount of errors in the reconstructed signal, the worse become the correlation results in terms of unwanted correlation artifacts. Therefore it is preferred to acquire a copy of the transmitted signal as clean as possible. The reference signal estimation capability can be improved by exploiting antenna diversity. In this paper we present the technique maximal ratio combining, which exploits the reception of multiple antennas to optimize the reference signal estimation. Here we also exploit both linear (vertical and horizontal) polarizations on receive to further improve the estimated reference signal. We show the improvement by using real data from an airborne passive radar system and by comparing the results of the antenna diversity approach to the results when using the reference signal from one antenna only.
11:20 Multi-carrier Adaptive Detection in Polarimetric Passive Radars
Francesca Filippini and Fabiola Colone (Sapienza University of Rome, Italy)
A polarimetric adaptive detection scheme based on a multichannel autoregressive process was recently proposed to improve the target detection capability of passive radar systems. Based on its demonstrated benefits, in this paper, the derived polarimetric detector is extended to the case of a passive radar simultaneously collecting signals transmitted by the same broadcast illuminator of opportunity on multiple carrier frequencies. Specifically, we derive a multi-frequency polarimetric adaptive detector that fruitfully leverages the information diversity conveyed in both the polarimetric and frequency domain. We demonstrate the effectiveness of the proposed strategy by means of an experimental validation against real data. To this aim, we make use of signals collected using an FM radio multichannel passive radar prototype, equipped with two dual-polarized antennas and able to collect signals at four frequency channels simultaneously. The application of the derived strategy also in comparison with state-of-the-art approaches shows the capability of the proposed technique to effectively increase the detection capability thanks to both a more effective rejection of the interference contribution at the single frequency channel as well as to an increased robustness with respect to time-varying characteristics of the employed waveform.
11:40 Joint Measurement of Target Angle and Angular Velocity Using Interferometric Radar with FM Waveforms
Jason Michael Merlo and Jeffrey Nanzer (Michigan State University, USA)
An architecture for a direct-downconversion complex correlation interferometric radar capable of the direct, joint measurement of angle and angular velocity of a point-target is presented. Due to the simplicity of the system, this technique can be implemented on existing radars with distributed receivers and a transmitter capable of frequency modulation. Derivations for the interferometric measurement of angle and angular velocity and guidelines for the design of such a system are presented. Simulations are provided for the proposed system with varying pulse-width, antenna baseline, and pulse center frequency. The results for all simulations show a low root-mean-square error which demonstrates the feasibility of the proposed system.

TM-L3: Phased Array Antennas

Room: Ch3
Chair: Xiangrong Wang (Beihang University, China)
10:20 Phased-Array Antenna Pattern Optimization with Evolution Strategies
Paul Dufossé (Inria & Ecole Polytechnique, Thales DMS, France); Cyrille Enderli (Thales Airborne Systems, France); Laurent Savy (THALES DMS, France); Nikolaus Hansen (Inria, France)
In this paper we address the problem of finding Phase-Only tapers configuration for a Phased-Array antenna that would be relevant for airborne radar applications. Following a constrained optimization approach, we apply the state-of-the- art blackbox optimization algorithm CMA-ES and build a novel constraint-aware fitness function based on rankings. To the best of our knowledge, this technique has never been used in constrained continuous optimization and proves to be robust as it provides satisfying results in a diversity of considered cases. Our approach is also convenient as it requires very few parameter tuning. We discuss the obtained results and how they could be extended.
10:40 Multi-channel Microwave-Photonic Link for Antenna Remoting in Multifunctional Phased-Array Radar
Luca Banchi (GEM Elettronica srl, Italy)
By combining microwave solid-state devices and photonic technologies an outperforming 50 Watts Transmitter and Receiver Module (TRM) with a very low Noise-Figure has been successfully realized at GEM Elettronica. The developed TRM, based on microwave-photonic high efficiency conversion, integrates all solid-state power amplifiers, limiter, circulator and optical components in an extremely compact design ideal for radar antenna remoting. A 16x 50 Watts microwave-photonic TRMs configuration has been developed and tested in order to realize an 800 Watts remote phased-array Antenna for high-performance 3D Multifunctional X-band radar.
11:00 Two-way MIMO Sparse Array Antenna Optimization with NSGA-III
Asgeir Nysaeter (FFI, Norway)
The paper uses the NSGA-III multiobjective optimization algorithm to optimize two-dimensional antenna locations with regard to MIMO two-way antenna patterns over a search grid. The optimization objectives are the mainlobe width and peak sidelobe level for predefined directions in the search area. For the receive problem the MIMO virtual array is optimized assuming transmission of orthogonal beams from the antenna. For the transmit problem semidefinite relaxation technique is used to minimize the difference between the antenna beam power and a desired pattern. An array antenna grid pattern of 40x40 grid points is assumed, and 64 TX/RX cards are used for the location optimization.
11:20 AESA Adaptive Beamforming Using Deep Learning
Simone Bianco (University of Milano Bicocca, Italy); Maurizio Feo (MBDA Missile Systems, Italy); Paolo Napoletano (University of Milan, Bicocca, Italy); Alberto Raimondi (University of Milano-Bicocca, Italy); Giovanni Petraglia and Pietro Vinetti (MBDA Missile Systems, Italy)
In this work we propose a method for the adaptive beam-forming of an antenna array using Deep Learning. The proposed method is based on a deep Convolutional Neural Network that takes as input an image-like radiation pattern encoding the desired behavior and computes the optimal currents needed to adapt the antenna to the new beam specification. The proposed approach drastically reduces the computation time (up to 1700×) introducing a smart mapping of a classic iterative algorithm to an antenna to reproduce it. After training the model is able to compute optimal currents successfully in a single forward pass, avoiding the need of expensive iterative optimizations to find the needed currents.
11:40 Co-existence of AESA (Active Electronically Scanned Array) Radar and Electronic Warfare (EW) Systems on Board of a Military Ship
Salvatore Celentano (Leonardo spa, Italy); Alfonso Farina (Leonardo Company Consultant, Italy); Luca Timmoneri (Leonardo Spa, Italy); Goffredo Foglia (Elettronica S.P.A., Italy)
In the context of the design of modern military ships, a significant issue is the Topside (or integrated mast) design i.e. the configuration of the equipment, sensors and actuators that constitute the Naval Combat System (SdC). The arrangement of the sensors affects the performance of the radio-frequency devices installed on board and therefore their effective co-existence. This paper reports on a simplified yet effective study to minimize the negative impact of a radar e.m radiation on an Electronic Warfare (EW) system when the two equipment are positioned on the same mast of a ship.

TM-L4: Radar Detection

Room: Ch4
Chairs: Stéphanie Bidon (University of Toulouse / ISAE, France), Danilo Orlando (Universita' degli Studi Niccolo' Cusano, Italy)
10:20 Subspace-Based Target Detection in the Presence of Multiple Alternative Hypotheses
Eloisa Faro (Università Roma 3, Italy); Gaetano Giunta (University of Roma Tre, Italy); Sudan Han (National Innovation Institute of Defense Techonology, China); Danilo Orlando (Universita' degli Studi Niccolo' Cusano, Italy); Luca Pallotta (University of Roma Tre, Italy)
This paper describes a new framework that, exploiting the Kullback-Leibler Divergence, allows to address the design of one-stage adaptive detectors for multiple hypothesis testing problems. Precisely, at the design stage, the problem is formulated in terms of multiple alternative hypotheses competing with the null hypothesis. Then, a one-stage decision scheme is derived in the context of both known model and unknown parameters as well as for the most general case of unknown model and parameters. Interestingly, the resulting detectors are given by the sum of the compressed log-likelihood ratio based on the available data and a penalty term depending on the number of unknown parameters. This general architecture is then particularized to the problem of subspace target detection, and its effectiveness is assessed through simulations also in comparison with its counterpart based on the two-stage paradigm.
10:40 Three-Dimensional Deterministic Detection and Estimation Algorithms for MIMO SFCW Radars
Alessandro Davoli, Emilio Sirignano and Giorgio M. Vitetta (University of Modena and Reggio Emilia, Italy)
In this paper, the problem of detection and joint estimation of range, azimuth and elevation of multiple targets in a multiple-input multiple-output stepped-frequency continuous wave radar system is investigated. Two iterative deterministic algorithms are illustrated and their accuracy is assessed on the basis of the raw data acquired from a commercial radar device. Our results evidence that these algorithms achieve similar accuracies, but at the price of substantially different computational efforts.
11:00 Improving Detection of Unknown Signal with Unknown Duration Using an Information Criterion
Abigael Taylor (ONERA, the French Aerospace Lab, France); Olivier Rabaste (Onera, France)
Detection using a noisy reference signal is a problem that arises frequently, in particular in passive radar applications. In this paper, we go further on the incorrect knowledge of the signal, as its time-support is supposed to be unknown - either too short or too long. We first show and illustrate that this additional error on the signal model deteriorates the detection performance. Two strategies to recover the time-support are proposed, the first one based on the Akaike Information Criterion, and the second one on optional additional constraints on the signal. The improvement in detection performance offered by these strategies is studied through numerical simulations.
11:20 Persymmetric Adaptive Detector for FDA-MIMO Radar
Cheng Jie, Hui Chen, Ronghua Gui, Jia Wenkai and Wen-Qin Wang (University of Electronic Science and Technology of China, China)
In this paper, under the condition of unknown Gaussian interference-noise covariance matrix, a persymmetric adaptive detector is proposed for frequency diverse array multi-input multi-output (FDA-MIMO) radar. With prior information of the interference-noise covariance matrix structure, an unbiased estimate is first obtained for the covariance matrix, followed by the analytical expression of detection probability. Numerical results show that, the proposed persymmetric adaptive detector for FDA-MIMO radar outperforms the conventional counterpart.
11:40 Hybrid Detection Approaches Using the Single Data Set Algorithm
Elias Aboutanios (University of New South Wales, Australia); Luke Rosenberg (Defence, Science and Technology Group & University of Adelaide, Australia)
The dynamic nature of sea clutter can lead to significant degradation of the performance of traditional coherent target detectors such as the Generalised Likelihood Ratio Test (GLRT) and Adaptive Matched Filter (AMF). These detectors are designed for complex Gaussian distributed clutter and rely on the availability of homogeneous training data. However, depending on the radar waveform and collection geometry, these conditions can be violated in maritime environments as the statistics of the clutter are slowly varying over time and range and can be non-Gaussian. The Single Data Set (SDS) algorithms were proposed to mitigate this problem by foregoing the need for training data. In this work we propose to combine the AMF and SDS approaches by devising a novel strategy for capitalising on both test and training data to improve the detection performance. Unlike heterogeneity screening strategies such as the Generalised Inner Product (GIP), the new technique is informed by the clutter statistics to employ all of the available range gates. The new algorithm performance is compared with the AMF, SDS and GIP-based hybrid detectors and shown to offer good performance in homogeneous environments and superior performance in heterogeneous environments.

TM-SS1: Urban remote sensing using SAR.

Room: Ch5
Chair: Paolo Gamba (Università degli Studi di Pavia, Italy)
10:20 Compressed Sensing-Based Multi-Aperture Focusing of Spaceborne Transmitter/Stationary Receiver Bistatic SAR Data
Adrian Focsa (University Politehnica of Bucharest / Military Technical Academy, Romania); Mihai Datcu (German Aerospace Center, Germany); Andrei Anghel (University Politehnica of Bucharest, Romania)
In this paper, we introduce a compressive sensing- based approach for increasing bistatic synthetic aperture radar imaging quality in the context of multi-aperture acquisition. The analyzed data is recorded over an opportunistic bistatic setup including a stationary ground based receiver (COBIS) and Sentinel 1 C-band transmitter. Since the terrain observation by progressive scans mode is operated, the receiver is able to record synchronization pulses and echoed signals from the scene during many apertures. Hence, it is possible to improve the azimuth resolution by exploiting the multi-aperture data. Obviously, the recorded data is not contiguous and a naive integration of the chopped azimuth phase history would gen- erate undesired grating lobes. The proposed processing scheme exploits the natural sparsity characterizing the illuminated scene. Therefore, a matching pursuit compressive sensing algorithm is employed for filling the "gaps" occurring in Doppler spectrum. The sparsifying basis/dictionary is constructed using the synthetic generated azimuth chirp. The obtained results show a significant improvement of the azimuth resolution along with dramatic attenuation of the side-lobes.
10:40 Generation of Large-Scale High Quality 3-D Urban Models
Yilei Shi (Technical University of Munich, Germany); Richard Bamler (German Aerospace Center (DLR), Germany); Yuanyuan Wang (Technical University of Munich & German Aerospace Center, Germany); Xiao Zhu (Technical University of Munich, Germany)
Interferometric synthetic aperture radar (InSAR) techniques are powerful tool for reconstructing the 3-D position of scatterers, especially for the urban areas. Since the estimation accuracy depends on the inverse of number of interferograms and signal-to-noise ratio (SNR), it is necessary to use as many as possible interferograms in order to achieve more accurate result. However, the number of interferograms of TanDEM-X data is generally limited for most areas. Therefore, in order to maintain the estimation accuracy, one feasible way is to increase the SNR. In this work, we proposed a novel framework, which integrates the non-local procedure into SAR tomography inversion and combines the robust estimation. A large-scale demonstration has been carried out with five TanDEM-X bistatic data, which covers the entire city of Munich, Germany. Quantitative evaluation of the reconstructed result with the LiDAR reference exhibits the standard deviation of the height difference is within two meters, which implies the proposed framework has great potential for high quality large-scale 3-D urban modeling.
11:00 Scattering Component Selection and Multi-resolution Detection of Persistent Scatterers in Sparsely Urbanized Areas
Gianfranco Fornaro (CNR-IREA, Italy); Antonio Pauciullo (IREA, CNR, Italy); Diego Reale and Simona Verde (CNR-IREA, Italy)
Approaches based on the coherent processing of Synthetic Aperture Radar (SAR) data archives acquired over long temporal intervals allows 3D reconstruction/imaging and monitoring of urban areas with very high accuracy on the velocity, e.g. sub-centimetric per year. The technology is rather mature and developed in some cases also at the level of operative services. Nevertheless research is still going on to improve the technology in terms of coverage and reliability of estimated parameters for correct localization and accurate monitoring of targets. With this regard, the application to regions characterized by a mixture of urban and rural areas is of interest. This work discusses the different techniques for filtering and selection of scattering mechanisms and for single and multi-look approaches for the multi-resolution detection of Persistent Scatterers.
11:20 Using Neural Networks for Change Detection and Classification of COSMO-SkyMed Images
Alessia Benedetti (University of Rome "Tor Vergata, Italy); Ludovica Porzio (Tor Vergata, Italy); Giovanni Schiavon (Tor Vergata University - DISP, Italy); Daniele Latini (University of Rome "Tor Vergata", Italy); Luca Fasano (Italian Space Agency, Italy); Fabio Del Frate (Università degli Studi di Roma, Italy)
In this paper a new approach based on the use of Synthetic Aperture Radar COSMO-SkyMed products to verify changes and to classify them is presented. Through an unsupervised neural network PCNN (Pulse Coupled Neural Network), the objects present in the images are analyzed and, eventually, the variations occurred between the two times analyzed. If the number of pixels changed is greater than a suitably threshold, the product is reclassified. The latter operation is based on a supervised neural network MLP (Multi Layer Perception) which, after a specific training phase, is able to produce the landcover map by taking as an input the backscattering coefficient and texture parameters of CSK products.
11:40 Bistatic Scattering from a Canonical Building
Gerardo Di Martino (Università di Napoli Federico II, Italy); Alessio Di Simone (University of Naples Federico II, Italy); Walter Fuscaldo (Consiglio Nazionale delle Ricerche (CNR), Italy); Antonio Iodice (Università di Napoli Federico II, Italy); Daniele Riccio (University of Naples Federico II, Italy); Giuseppe Ruello (Università di Napoli Federico II, Italy)
In this paper we present an analytical model for the evaluation of the electromagnetic (EM) scattering from a typical composite target of urban areas in a generic bistatic configuration. The considered scene comprises a canonical building modeled as a parallelepiped lying over a rough terrain. Closed-form expressions for the scattered EM field are derived under the framework of the Kirchhoff approximation, with subsequent Geometrical Optics and Physical Optics approximations adopted for evaluating scattering from building walls and ground, respectively. Single- and multiple-scattering contributions are properly modeled as well as the EM and geometric characteristics of both the composite target and the sensors, such as material composition, size, terrain roughness, building orientation, viewing angle, polarization, frequency. The proposed scattering model can be used in information retrieval procedures using bistatic radar data.

Tuesday, September 22 12:00 - 2:15 (Europe/Rome)

TM-P1: PSS1-Resource management for cognitive radar systems

Chairs: Junkun Yan (Xidian University, China), Wei Yi (University of Electronic Science and Technology of China, China)
Joint Optimization of Target Assignment and Resource Allocation for Multi-Target Tracking in Phased Array Radar Network
Chenguang Shi, Lintao Ding and Wei Qiu (Nanjing University of Aeronautics and Astronautics, China); Fei Wang (Nanjing University of Aeronautics and Astronautics, Nanjing, China); Jianjiang Zhou (Nanjing University of Aeronautics and Astronautics, China)
This study proposes a joint optimization scheme of target assignment and resource allocation for multiple target tracking in phased array radar network. The key idea of the proposed scheme is to simultaneously improve the multi-target tracking accuracy and minimize the total dwell time of the phased array radar network by optimizing the target-to-radar assignment, revisit time control, bandwidth and dwell time allocation, such that both the target tracking performance and the low probability of intercept (LPI) performance of the network system can be enhanced. The closed-form expression for the predicted Bayesian Cram\'{e}r-Rao lower bound (BCRLB) with target-to-radar assignment, revisit time, bandwidth and dwell time is derived, which is used to represent the tracking accuracies of multiple targets. The resulting mixed-integer, non-linear and non-convex optimization problem is subsequently solved by a developed efficient three-stage solution technique. Finally, simulation results demonstrate the advantages of the proposed scheme, in terms of the multi-target tracking accuracy and the achieved LPI performance of the underlying system.
Joint Online Route Planning and Power Allocation for Multitarget Tracking in Airborne Radar Systems
Xiujuan Lu (University of Electronic Science and Technology of China, China); Lingjiang Kong (University of Electronic Science and Technology of China (UESTC), China); Jun Sun and Ye Yuan (University of Electronic Science and Technology of China, China)
Reasonable route planning and power allocation strategy for the airborne radar systems can sufficiently utilize the limited power and improve the multi-target tracking (MTT) performance. However, in the existing strategies, the route planning and power allocation are usually considered separately and the practical online requirements of route planning have not been addressed. In view of it, we propose a joint online route planning and power allocation (JORPPA) scheme for the task of the MTT in the airborne radar systems in this paper. The posterior Cram\'{e}r-Rao lower bound (PCRLB) for the moving airborne system is derived and utilized as a constituent of the MTT performance metric. Then, the JORPPA algorithm is modeled as an optimization problem by using a scaled accuracy-based utility function as the objective function. The formulated problem is non-convex. We then propose a partition-based approach to solve it effectively. Numerical results verify the effectiveness of the proposed solution for MTT.
Antenna Placement Algorithm for Distributed MIMO Radar with Distance Constrains
Yao Wang, Wei Yi and Shixing Yang (University of Electronic Science and Technology of China, China); Mahendra Mallick (Independent Consultant, USA); Lingjiang Kong (University of Electronic Science and Technology of China, China)
This paper considers the optimal antenna placement problem for distributed multi-input multi-output (MIMO) radar system. The aim is to maximize the surveillance performance of the radar system in a certain interested area by adjusting the positions of radar antennas, while satisfying both the maximum and minimum distance constraints between each antenna pair. To evaluate the surveillance performance in a certain area, we use a coverage ratio metric as the optimization objective function. However, the formulated optimization problem is computationally intractable for practical scenarios due to its high dimensionality, non-convexity and especially the complex spatial constraints. To solve this problem, we further propose an enhanced particle swarm optimization (PSO) algorithm, which differs from the standard PSO algorithm in that its particles can properly take account into constraints during the swarm optimization process. Finally, numerical results are provided to verify the superior performance of the proposed antenna placement algorithm.
Sensor Selection for Multi-target Tracking in Phased Array Radar Network Under Hostile Environment
Jinhui Dai and Junkun Yan (Xidian University, China); Shenghua Zhou (National Laboratory of Radar Signal Processing, Xidian University, China); Wang Penghui and Bo Jiu (Xidian University, China); Hongwei Liu (National Laboratory of Radar Signal Processing, China)
In this paper, a novel sensor selection scheme is presented for multi-target tracking (MTT) in phased array radar (PAR) network under hostile environment. We consider a scenario that the active PARs transmit beams to track multi-targets, meanwhile, the enemy targets equipped with passive radars can collaboratively locate positions of PARs according to the intercepted signals. The main idea of this scheme is to appropriately adjust the beams of multiple PARs, with the aim of improving their MTT performance while decreasing the tracking capability of the enemy targets. We formulate the sensor selection scheme as a non-convex optimization problem, and solve it through convex relaxation and a proposed two-step solution technique. Simulation results demonstrate the superior performance of the proposed method when compared with the traditional random beam allocation scheme.
Mobile Sensors Path Planning for Cooperative Monitoring of Different Mission Importance Areas
Yaoying Tang and Yao Wang (University of Electronic Science and Technology of China, China)
This paper considers an offline path planning problem about cooperative monitoring in different mission importance areas of interest (AOIs) by mobile sensor nodes. Each node is equipped with a radar sensor and limited by maneuver constraints. To evaluate the cooperative surveillance performance, we propose an effective monitoring metric and a revisit metric, which are based on the detection performance of the mobile surveillance system in AOIs during the mission time. Since the formulated optimization problem is high dimensionality and has complex constraints, we propose a solution that contains two parts: 1) an optimization algorithm based on particle swarm optimization (PSO) to reduce the computation load with the constrains satisfied and 2) a path smoothing method based on third-degree B-spline to smooth the optimized paths. Finally, simulation results verify the feasibility and efficiency of the algorithm.
Joint Waveform Selection and Time-Space Resource Management in Netted Colocated MIMO Radar System for Multi-target Tracking
Yang Su (University of Electronic Science and Technology of China, China); Ting Cheng (University of Electronic Science and Technology, China); Zi-Shu He (University of Electronics Science and Technology of China, China); Xi Li (University of Electronic Science and Technology of China, China)
Netted colocated multiple-input multiple-output (C-MIMO) radar system inherits the advantages from the netted radar system and the C-MIMO radar, therefore, it has great research value. In this paper, the joint waveform selection and time-space resource management optimization model for the netted C-MIMO radar system (NCMRS) in multi-target tracking (MTT) is proposed, where the network system resource consumption and the overall tracking performance of multiple targets are taken into consideration comprehensively. Based on the model, a joint waveform selection and time-space resource allocation (JWSTSRA) algorithm is put forward, where the system sampling period, the activated nodes, the node-target assignment, and the sub-array number, transmit energy and transmit waveform of each activated node can be adjusted adaptively. Simulation results demonstrate the effectiveness of the proposed JWSTSRA algorithm.
Fast Constant Modulus MIMO Radar Waveform Design for Interference Mitigation
Hao Zheng (XiDian University, China); Bo Jiu (Xidian University, China); Hongwei Liu (National Laboratory of Radar Signal Processing, China)
In this paper, a fast constant modulus multiple-input multiple-output (MIMO) radar waveform design method is proposed for interference mitigation. In order to reduce the computational burden of optimizing waveform, a new transmit scheme is presented, which can decouple the spatial part and temporal part of the waveform design. Focusing on the spatial property of waveform, a novel cost function is proposed to minimize the maximum transmit power difference between interference region and target region. With the constant modulus constraint, the proposed cost function is intractable. To further improve the computational efficiency, a two-step strategy is proposed to obtain a suboptimal solution, which decomposes the original problem into a set of parallel beamformers design and a linear programming (LP) problem about the support time of beamformers. To optimize the beamformers efficiently, the majorization-minimization (MM) algorithm and its acceleration version are employed to tackle this problem. Numerical results show the efficiency of the proposed method.

TM-P2: Multistatic and passive radars

Chairs: Daniel O Hagan (Fraunhofer FHR, Germany), Krzysztof (Chris) Kulpa (Warsaw University of Technology, Spain)
Reinforcement Learning Based Dynamic Task Scheduling for Multifunction Radar Network
Longxiao Xu and Tianxian Zhang (University of Electronic Science and Technology of China, China)
In this paper, in order to improve the efficiency of tasks execution of multifunction radar network, a dynamic tasks scheduling problem is investigated. Considering the uncertainty of dynamic tasks request times, a dynamic tasks method that based on reinforcement learning is proposed. Firstly, we construct a Markov Decision Process (MDP) for the multifunction radar network executing tasks, and choose the dropped ratio of tasks as the evaluation criterion. Then, a model-free reinforcement learning framework for multifunction radar network executing tasks is formulated. Under the framework, we design the action space for this reinforcement learning question, and a method of tasks scheduling based on Q-learning is proposed. Finally, simulation results are provided to verify the validity of proposed method.
Airplane Detection by FSR Using Cosmic Radio Emissions
Hristo A. Kabakchiev (Sofia University "St. Kliment Ohridski", Bulgaria); Vara Behar (Institute of Information and Communication Technologies, Bulgaria); Ivan Garvanov (University of Library Studies and Information Technologies, Bulgaria); Dorina Kabakchieva (University of National and World Economy, Bulgaria); Avgust Kabakchiev (BULATSA, Bulgaria); Nikola Petrov (National Astronomical Observatory Rozhen Smolyan, Bulgaria); Hermann Rohling (Technical University Hamburg-Harburg, Germany); Mark Bentum (Eindhoven University of Technology & ASTRON, The Netherlands)
The paper analyses and compares the possibility for airplanes detection by a FSR system that exploits radio emission from such cosmic bodies as Moon, Sun and pulsars. This possibility is estimated as a magnitude of SNR at the input of the signal detector calculated depending on the frequency of reception, the size of airplanes, which are crossing the baseline of FSR at the approximately right angle and the distance from the receiver to airplanes.
Exploitation of Bi-Static Radar Architectures for LEO Space Debris Surveying and Tracking: The BIRALES/BIRALET Project
Angelo Podda (Vitrociset spa, Italy); Sergio Casu (Vitrociset Spa, Italy); Antonio Coppola (Vitrociset, Italy); Fabio Protopapa (Vitrociset spa, Italy); Andrea Lazzareschi Sergiusti (Vitrociset Spa, Italy); Germano Bianchi, Claudio Bortolotti, Mauro Roma, Giuseppe Pupillo, Luca Lama, Federico Perini, Marco Schiaffino, Andrea Maccaferri, Giovanni Naldi and Andrea Mattana (INAF - IRA, Italy); Tonino Pisanu and Enrico Urru (INAF - OAC, Italy); Luca Schirru (National Institute for Astrophysics - Astronomical Observatory of Cagliari, Italy); Pierluigi Ortu and Francesco Gaudiomonte (INAF - OAC, Italy); Pierluigi Di Lizia, Giovanni Purpura and Mauro Massari (Politecnico di Milano, Italy)
The space debris population is continuously growing and it represents a potential issue for spacecraft. New collisions could exponentially rise the amount of debris and so the level of risk represented by these objects. The monitoring of space environment is necessary to prevent new collisions. For this reason, radar measurements are relevant, in particular to observe objects in Low Earth Orbit. Regarding the Italian contribution, there are two radars based on two different radio telescopes as receivers: the BIRALES and the BIRALET systems. We propose a detailed description of these systems, focusing on hardware and software components that permit to perform range and range rate measurement of resident space objects.
Direct Localization of Shortwave Emitters in Multipath Ionosphere Channel
Lichan Yan, Xing-peng Mao and Minqiu Chen (Harbin Institute of Technology, China)
The conventional techniques designed for localization of shortwave emitters are two-step methods. Since the internal constraint of the received data is ignored, the two-step methods are considered suboptimal. Besides, an extra procedure of data association is required for the localization of multiple emitters. To overcome the drawbacks mentioned above, methods which can directly determine the locations of emitters attract attention of the researchers. However, virtual height of reflection in an unknown ionosphere is hard to estimate, the relation between the location of emitter and the steering vector cannot be determined. In such scenario, the direct localization is no longer work. To overcome this limitation, we present an ionosphere multipath reflection model. In this work, a novel method is proposed to decouple angles. By taking advantages of the decoupled technique, the direct localization methods do not require multi-dimensional search caused by unknown ionosphere virtual heights, and can decrease the computational complexity. Simulation results verify the effectiveness of the proposed method.
An Adaptive Optimization Algorithm for Antenna Deployment in Dynamic Environment
Ziqin Wang, Yanli Zeng, MingMin Shu, Mingfeng Pu and Chuanlin Huang (Science and Technology on Electronic Information Control Laboratory, China)
In this paper, an adaptive convergence method is investigated to optimize the prediction-based particle swarm optimization (PBPSO) algorithm that has been adopted for multistatic radar system surveilling multiple dynamic regions. We propose two parameters to measure the convergence and diversity. Unlike the traditional PBPSO method, which requires extensive experimentation to determine the number of iterations, the proposed method measures the Pareto Front's convergence degree and diversity to stop the iteration adaptively resulting in significantly better efficiency solving the optimal problem. Firstly, by calculating the parameter proposed in this paper, the algorithm perceives its degree of convergence and diversity. Then by comparing its own degree of convergence and diversity with a threshold set in accordance with the input requirements, the algorithm can determine whether to stop iteration adaptively. Numerical results show the proposed algorithm outperforms the traditional one in optimizing the antenna deployment scheme in a dynamic environment.
Topology Design for Distributed Radar Imaging Based on Wavenumber Domain Splicing Analysis
Fanyun Xu, Rufei Wang, Junyu Zhu, Yongchao Zhang, Yin Zhang and Yulin Huang (University of Electronic Science and Technology of China, China); Jianyu Yang (School of Electronic Engineering, China)
Distributed radar is more flexible because of transmitters and receivers are separately distributed in this system. It can form multiple bistatic SAR systems according to the detection requirements in a self-organized manner, and obtain more effective information in a shorter time. During the movement of the radar platform, a specific spatial spectrum will be generated in the wavenumber domain. By splicing the spatial spectrum of multiple bistatic SAR systems, the resolution in the range direction or cross range direction can be improved. In this paper, the rule of spatial spectrum distribution in the wavenumber domain is analyzed in detail, and a method is provided for the topology design of radar platforms in combination with the splicing spatial spectrum. Numerical simulation gives imaging results after topology design, which demonstrates the effectiveness of distributed radar topology design method for improving range or cross range resolution.
Radiometric Passive Imaging for Robust Concealed Object Identification
San Hlaing Myint (Global Information and Telecommunication Institute Waseda University); Yutaka Katsuyama (Global Information and Telecommunication Institute Waseda University, Japan); Toshio Sato (Waseda University, Japan); Xin Qi (Global Information and Telecommunication Institute, Waseda University, Japan); Zheng Wen, Keping Yu, Kiyohito Tokuda and Takuro Sato (Waseda University, Japan)
Artificial Intelligence (AI) based millimeter wave radiometric imaging has become popular in a wide range of public security check systems, such as concealed object detection and identification. However, the low radiometric temperature contrast between small objects and low sensitivity is restricted to some extent. In this paper, an advanced radiometric passive imaging simulation model is proposed to improve the radiometric temperature contrast. This model considers additional noise, such as blur, variation in sensors, noise sources and summation of the number of frames. We establish a comprehensive training dataset that considers the physical characteristics of concealed objects. It can effectively fill the lack of a large database to avoid deteriorating the identification accuracy of AI applications. Moreover, it is also a key solution for improving the robustness of AI based object identification by using a convolutional neural network (CNN). Finally, simulation results are presented and analyzed to validate the proposed comprehensive training dataset and simulation model. Consequently, the proposed simulation model can effectively improve the robustness and accuracy of AI-based concealed object identification.
Iso-Motion-Compensation Curves for Bistatic SAR
Yi Li, Wenchao Li, MIn Li and Wenjing Wang (University of Electronic Science and Technology of China, China); Zhongyu Li and Junjie Wu (University of Electronic Science and Technology of China (UESTC), China); Yulin Huang and Jianyu Yang (University of Electronic Science and Technology of China, China)
Motion error and compensation is a crucial problem for bistatic SAR. In this paper, based on the geometry model of bistatic SAR considering motion error and topographic variation, the concept of bistatic SAR iso-MoCo curve is proposed, which refers to the contour line of motion error compensation for bistatic SAR. Analytical expression of bistatic SAR iso-MoCo curves under flat terrain conditions is deduced via a conic section model. Properties of motion errors considering spatial variation with and without topographic variation are analysed based on simulation results.
2D Constrained PBR Localization via Active Radar Designation
Angela Marino (University of Naples Federico II, Italy); Augusto Aubry (Universita degli studi di Napoli, Italy); Antonio De Maio (University of Naples "Federico II", Italy); Paolo Braca (CMRE, Italy)
A new algorithm for Passive Bistatic Radar localization is proposed capitalizing measurements gathered by a co-located active radar and exploiting multiple illuminators of opportunity. Prior information, related to the Passive Bistatic Radar (PBR) receive antenna main-beam size and to the uncertainty characterizing active radar data, is accounted for formalizing ad-hoc constraints for the localization process. Hence, the estimation task is cast as an elliptic positioning problem, according to the constrained Least Squares (LS) framework. The performance of the proposed estimator exhibits sensible improvement with respect to some counterparts.
A Novel Phase Compensation of the Target Detection for CSI with Commodity WiFi
Xiang Zhu, Bin Zhao, Zhang Yun, Jin Wei and Hongbo Li (Harbin Institute of Technology, China)
This paper aims at discussing the potential solutions of detecting and sensing based on WiFi signal in indoor environment. The channel state information (CSI) is a good choice to work as a new kind of radar pulse signal. However the differences between the communication receiver and radar's is a big challenge we must overcome, such as unexpected phase rotation. According to wireless local area network (WLAN) protocol, A lot of pilot signals are scattered in the data field which may help us. Finally, a method was introduced to achieve joint sensing.
Network Architecture Optimization for Area Surveillance in Multi-static Radar Systems
Chengxin Yang, Yao Wang and Wei Yi (University of Electronic Science and Technology of China, China)
This paper describes a network architecture optimization method for area surveillance in multistatic radar systems. The proposed method addresses the problem of contradiction between the network performance, and the limited communication and computing ability of the multistatic radar systems. The goal is to improve the system surveillance performance by optimizing the network architecture of such systems. We take the coverage ratio of the surveillance area as the optimization index, consider the communication and calculation abilities of radar nodes as constraints, and establish an optimization model for the network architecture. Two methods of data processing for detection are considered to enhance the node communication and computation abilities. It is shown that the optimization problem is nonconvex, and a clustering-based sequential optimization algorithm is proposed as a sub-optimal solution method. Finally, simulation results verify the validity and effectiveness of the proposed method.
Efficient Direct Signal Cancellation for FM-based Passive Radar
Christof Schüpbach (Armasuisse Science+Technology, Switzerland); Stephen Paine (Armasuisse Science+Technology, Switzerland & University of Cape Town, South Africa); Daniel O'Hagan (University of Cape Town, South Africa)
The performance of an FM-based passive radar sensor is highly dependant on the evel of suppression of direct signal interference. In this paper, we compare the performance of a selection of suppression algorithms in terms of detection performance and processing time. We show that the ECA-CD algorithm, even though originally designed for OFDM signals, can be applied to FM signals and that it performs well while being very computationally efficient.

TM-P3: Wideband antennas and phased array

Chair: Seyed Mohammad Karbasi (Sharif University of Technology, Iran)
Research on the Effect of Phased Array Beam Scanning on Self-interference Cancellation
Jie Zhang (NRIET, China); Wensheng Chang (Nanjing Research Institute of Electronics Technology, China)
Aiming at the application of full-duplex (FD) method in phased array system, based on the modeling of coupling path characteristics for different spatial links using adaptive system identification method, the effect of self-interference cancellation (SIC) influence of beam scanning toward different directions is studied under practical experiments. From the experiments' results, the performance difference of self-interference cancellation after adaptive filtering for different beam scanning directions between the best result and the worst is more about 33dB. As for the boresight line of the transmission array, the SIC result is the best. If the scanning directions are off the boresight line, the SIC result is usually worse than that of the boresight line. Therefore, we can draw the conclusion that, for phased array system, beam scanning plays an important role on self-interference cancellation for simultaneous transmit and receive (STAR).
Modeling and Experimental Study of Full-Duplex Channel Characteristics for Phased Array Simultaneous Transmission and Reception
Jie Zhang (NRIET, China); Wensheng Chang (Nanjing Research Institute of Electronics Technology, China)
Aiming at the application of full-duplex (FD) method in phased array, the coupling path characteristics for different spatial links are modeled by adaptive system identification method based on practical experiments. From the experiments' results, the conclusions that, the thousands of coupling paths among the phased array need to be modeled separately, and the self-interference cancellation for different coupling paths must be dealt with using individual coefficients, can be reached. At the same time, the normalized least mean square (NLMS) algorithm is used to adaptively cancel the coupling interference. The results show that the interference can be cancelled about 38dB for wideband signal of 300MHz bandwidth, and about 68dB for single-frequency signal. For broadband channel feature modeling, more order adaptive filter systems are needed. In addition, the channel characteristics during phased array scanning are also analysed.
Research on Modeling and Principle Verification of Full-Duplex Technology Based on Phased Array
Jie Zhang (NRIET, China); Wensheng Chang (Nanjing Research Institute of Electronics Technology, China)
For full-duplex applications of phased array systems, the self-interference/mutual interference coupling link has complex characteristics and high self-interference power. The transmission and reception characteristics of a full-duplex channel were studied by using a phased array. In the integrated spatial beamforming interference suppression, RF interference cancellation and digital adaptive filtering processing, a multi-level comprehensive architecture, method and model of RF interference signals based on phased array system is proposed. Through principle test and simulation verification, the signal-to-interference and noise ratio is improved by 13dB for the reception matching processing of the desired signal. Therefore, in full-duplex applications of phased array systems, this method and the corresponding optimization model can effectively reduce the power of RF interference signals to support the simultaneous application of multiple functions.
Design of a Spherical Conformal Phased Array Antenna Based on the Truncated Icosahedron
Jihong Zhang, Peiguo Liu, Zhaowen Zhuang, Jibo Wei, Xiaohui Liu, Hongqiang Wang and Ligang Li (National University of Defense Technology, China)
Spherical phased array antenna is widely used in satellite communication, weather surveillance and aircraft detection thanks to its hemispherical scanning potentials. In this paper, we designed a spherical conformal phased array antenna through pentagon and hexagon sub-arrays that following the truncated icosahedron configuration. The radiator inside each sub-array was an array of dual-polarized microstrip antennas. Full polarization is thereby formed due to the superimposition of these sub-arrays with various orientations. For a targeted direction, the active sub-arrays and antennas were defined first. Then pattern synthesis simulations were conducted by convex optimization methods to obtain the excitations for desired patterns. It is demonstrated that this antenna array could achieve hemispherical coverage while reducing the fabrication cost and facilitating the pattern synthesis process.
Robust Adaptive Beamformer Based on Weighted Sparse Constraint
Qin He, Ziyang Cheng, Zhihang Wang and Zishu He (University of Electronic Science and Technology of China, China)
Considering the high sidelobe of the MVDR beamformer, a sparse constraint on beam pattern has been suggested to suppress the sidelobe of adaptive beamformer. Based on the sparse constraint, a modified Capon spectrum is proposed to form a new beamformer to compensate for the lack of robustness of the MVDR beamformer. In proposed method, the constructed penalty matrix gives the minimum penalty on the mainlobe region to keep its shape in existing steering vector error and meanwhile the positions of interferences distinguished from the sidelobe region by the modified Capon spectrum are imposed a larger penalty to obtain a performance of interference suppression as perfect as the MVDR beamformer's. What's more, the sidelobe level in our method is moderately controllable. Finally, an algorithm of the alternating direction method of multipliers (ADMM) is suggested to solve the problem efficiently. Numerical examples show the advantages of the proposed method in adaptive beamforming.
Difference Pattern Synthesis for Spherical Arrays
Zhijiang Huang (China Academy of Engineering Physics, China); Yimao Sun (China Academy of Engineering Physics & University of Electronic Science and Technology of China, China)
Spherical arrays with array processing based on spherical harmonics have been studied for a wide range of applications in the smart antenna community and the acoustic array community. In this paper, a direct difference pattern synthesis method for spherical array based on spherical harmonics is proposed. We first propose to synthesize the spherical array difference pattern with spherical harmonics of degree ±1. Then the mapping relationship between the spherical array difference pattern and the linear array difference pattern is studied with the help of Legendre functions decomposition. With this mapping, uniform linear array specific difference pattern synthesis techniques become readily available in closed form for spherical arrays.
Performance Analysis and Evaluation of Implementing the MVDR Beamformer for the Circular Antenna Array
Somayeh Komeylian (Ryerson University, Canada)
In the area of pattern array synthesis, highly-directional radiation pattern guarantees accuracy and resolution, [1], which are accompanied by a drastic reduction in SLL. Pattern synthesis techniques are characterized by the two distinct scenarios; (1) the design of antenna array geometries for steering beampattern in an arbitrary direction in the space, and (2) the implementation of beamforming techniques for steering beampattern in the direction of interest in the space. This study has a major contribution for fully evaluating the performance of implementing the minimum variance distortionless response (MVDR) beamformer for the circular antenna array geometry in comparison with the different linear array geometries. A full and quantitative comparison between the circular antenna array and the different available linear antenna array geometries have been rigorously fulfilled for highlighting differences and advantages of the performance of the circular antenna array geometry over the different linear array antenna geometries using the four following concepts of (1) spatial correlation function (SCF), (2) efficiency, (3) signal to interference ratio (SIR), and (4) propagation time delay.
Design of an HF Transmitter Antenna for Bistatic Ionospheric Soundings in Antarctica
Kathleen C MacWilliam (University of Cape Town, South Africa)
This paper presents the design of a suitable high frequency (HF) transmitter antenna for installation at the South Pole as part of a low-powered bistatic ionospheric sounding system intended to detect traveling ionospheric disturbances (TIDs). A highly directional antenna was required so as to reduce interference with a nearby South Pole SuperDARN radar. HF ionospheric propagation was investigated, with the polar ionosphere and its impact on system functionality being of particular concern. Freely available propagation prediction tools were reviewed and ICEPAC was selected for use based on its high-latitude capabilities. Manual calculations of the non-deviative radio wave absorption in the lower layers of the ionosphere were done (for both extraordinary and ordinary wave modes) by using the magnetoionic Appleton-Hartree equations in conjunction with relevant ionospheric and geophysical models. These results were used to supplement transmission losses estimated with ICEPAC so as to ensure that enough power is supplied to allow for both wave modes to reach the receiver. The properties of the lossy ground at the South Pole were researched and a multi-layered substrate ground plane was modeled for use in FEKO simulations. Several antennas were investigated through an iterative design process and a three element rectangular loop Yagi-Uda was chosen for final consideration.

TM-P4: Cognitive radars

Chairs: Alfonso Farina (Leonardo Company Consultant, Italy), Sevgi Z Zubeyde Gurbuz (University of Alabama & TUBITAK Space Technologies Research Institute, Italy)
Deinterleaving Pulse Trains via Interleaved Markov Process Estimation
Gabriel Ford, Benjamin Foster and Stephen Braun (Lockheed Martin Advanced Technology Laboratories, USA)
A pulse train deinterleaving framework based on interleaved Markov process (IMP) estimation is presented. The approach models interleaved pulse trains as a collection of Markov processes interleaved by a Markov switch. Deinterleaving is performed through learning an IMP representation of the interleaved pulse trains. We apply a recent unsupervised information-theoretic technique that infers the underlying Markov processes by minimizing a penalized maximum likelihood (PML) entropy cost function. The technique is able to associate groups of disparate clusters in the pulse parameter space that are produced by the same parameter-agile emitter. This enables the correct determination of the number of emitters in scenarios where traditional multi-parameter clustering techniques produce many more clusters than there are actual emitters. Pulse timing information is incorporated to the extent that it is reflected in pulse ordering; in constrast to traditional time-of-arrival analysis techniques, the approach does not require precise timing resolution and can accommodate highly complex pulse repetition interval (PRI) patterns. An experimental demonstration and performance characterization on synthetic pulse train datasets is provided.
Fully Adaptive Resource Management in Radar Networks
Roland Oechslin (Armasuisse, Switzerland); Sebastian Wieland (Büro für Sensorik und Signalverarbeitung, Germany); Andreas Zutter (PrecisionWave AG, Switzerland); Uwe Aulenbacher (Büro für Sensorik und Signalverarbeitung, Germany); Peter Wellig (Armasuisse, Switzerland)
We present a fully adaptive radar network testbed that consists of four monostatic X-band radar sensors and a centralized controller which uses a perception-action cycle to optimize radar parameter in real time. The sensor hardware and the control algorithms are presented and explained. In a second part of the paper, we present results from outdoor experiments where an optimized radar resource allocation has been considered. Given a pre-defined track performance, the radar network is able to save resources by reducing the bandwidth to the amount needed for an accurate tracking and by operating only the sensors with the best measurement contributions to the active tracks.
Radar-camera Fusion for Road Target Classification
Kheireddine Aziz (Interuniversity Microelectronics Centre (IMEC) & Vrije Universiteit Brussel (VUB), Belgium); Eddy De Greef, Maxim Rykunov and Andre Bourdoux (IMEC, Belgium); Hichem Sahli (VUB, Belgium)
This paper presents a radar and camera sensor fusion framework as a vulnerable road user (VRU) perception system that can automatically detect, track and classify different targets on road. The first module of the system performs a spatial- temporal alignment on a common plane of detections provided by the radar signal processing and video processing modules. The second module is dedicated to data association of the aligned detections. A centralized fusion algorithm takes the current aligned detection set (locations and labels) as inputs from both sensors and performs multi-object tracking with a joint probabilistic data association (JPDAF) algorithm underlying the Kalman filter. The proposed radar/camera fusion system is experimentally evaluated through multi-object tracking scenarios. The experimental results demonstrate its reliability and effectiveness compared to a single sensor system.
Multistatic Radar Deployment Within A Non-Connected Region
Yi Han, Tianxian Zhang and Xiaobo Yang (University of Electronic Science and Technology of China, China)
In this paper, an optimal multistatic radar deployment problem is studied under the assumption of the nonconnected deployment region. By dividing the non-connected deployment region as an union of multiple connected subregions, the deployment problem can be modeled as a mix-integer nonlinear programming problem (MINP). The optimization variables of this deployment problem consist of selection of subregion and location optimization within every subregion. To solved this deployment problem, we first construct the mathematical optimization model, whose optimization objectives include the effective coverage and distribution uniformity of radar. Then, to alleviate the shortage of solving MINP in the conventional method, a pertinent multi-objective particle swarm optimization (MOPSO) variant for MINP is developed by modifying the dynamics of the particle motion for MOPSO. Finally, numerical results are provided to verify the validity of the proposed algorithm.
Cognitive Radar System for Obstacle Avoidance Using In-Motion Memory-Aided Mapping
Liyong Guo (Beijing Institute of Technology, China); Michail Antoniou and Christopher J. Baker (University of Birmingham, United Kingdom (Great Britain))
The paper introduces a radar signal processing method for goal-oriented, collision-free navigation in mobile robotic platforms. The derived algorithm creates an enhanced perception of the area in front of the sensor through accumulating a sequence of radar pulses that is constantly updated, and uses previously obtained perception to inform future robot steering actions on the fly, thus creating a form of working memory. The algorithm is analytically described, and experimentally confirmed in laboratory conditions with a ground mobile robot operating in real-time.
AI-Augmented Multi Function Radar Engineering with Digital Twin: Towards Proactivity
Mathieu Klein (Thales Air Systems, France); Thomas Carpentier (THALES LAnd & Air Systems, France); Eric Jeanclaude (THALES Land & Air Systems, France); Rami Kassab (Thales Air Systems, France); Konstantinos Varelas and Nico de Bruijn (THALES Land & Air Systems, France); Frederic Barbaresco (Thales Air Systems, France); Yann Briheche (THALES Research & Technology, France); Yann Semet (THALES Group, Germany); Florence Aligne (Thales Research and Technology, France)
Thales new generation digital multi-missions radars, fully-digital and software-defined, like the Sea Fire and Ground Fire radars, benefit from a considerable increase of accessible degrees of freedoms to optimally design their operational modes. To effectively leverage these design choices and turn them into operational capabilities, it is necessary to develop new engineering tools, using artificial intelligence. Innovative optimization algorithms in the discrete and continuous domains, coupled with a radar Digital Twins, allowed construction of a generic tool for "search" mode design (beam synthesis, waveform and volume grid) compliant with the available radar time budget. The high computation speeds of these algorithms suggest tool application in a "Proactive Radar" configuration, which would dynamically propose to the operator, operational modes better adapted to environment, threats and the equipment failure conditions.
Generating Synthetic Short-Range FMCW Range-Doppler Maps Using Generative Adversarial Networks and Deep Convolutional Autoencoders
Marcio L Lima de Oliveira (University of Twente, The Netherlands); Marco Bekooij (University of Twente and NXP Semiconductors, The Netherlands)
In this paper, we discuss the usage of Generative Adversarial Networks (GANs) and Deep Convolutional Autoencoders (CAE) for creating synthetic Range-Doppler (RD) maps of Frequency-Modulated Continuous-Wave (FMCW) radars for a short-range situation with moving objects, based on measured RD maps of pedestrians and cyclists. Instead of using regular mathematical functions or heavy radar simulations, we have used an Artificial Neural Network (ANN) model to generate new data. By using our synthetic data, we can automatically have ground-truth data without the need for manual labor; easily create large synthetic datasets; hardly use much computational power after training. To evaluate our method, we have trained a detector system with just synthetic data, and it was capable of detecting moving objects correctly, on actual Range-Doppler maps, 11.6% better than when using a small dataset.
Transmit-Receive Design for Non-Uniform Pulse Repetition Interval Airborne Radar in the Presence of Signal-Dependent Clutter
Tao Fan, Mengmeng Ge, Na Gan and Yanqin Xu (University of Electronic Science and Technology of China, China); Guolong Cui (University of Electronic Science and Technology of China (UESTC), China); Zhihao Jiang (Naval Research Academy, China)
This paper deals with the joint design of transmit waveform and receive filter to improve the clutter rejection capability for non-uniform pulse repetition interval (NUPRI) airborne radar. Specifically, a multipulse echo model accounting for a moving point-like target and signal-dependent clutter is first established. Then the echo is processed via the matching filter and the windowed Doppler filter bank to obtain the two-dimensional range-Doppler plane. Further, the integrated sidelobe level of clutter (ISLC) that spreads the region of the target of interest in the plane is considered to minimize forcing constant modulus constraint on the waveform. To solve the resultant non-convex problem, the Sequential Greedy Optimization Algorithm (SGOA) through alternately updating the receive filter and transmit waveform is proposed to monotonically decrease ISLC to converge. In each iteration, the iterative algorithm based on coordinate descent (CD) framework and the sequential convex approximation algorithm are, respectively, explored to obtain the transmit waveform and receive filter. Finally, the performance of the proposed algorithm is assessed through numerical simulations showing its capability to suppress signal-dependent clutter.

TM-P5: Target detection and clutter suppression

Chairs: Warren P du Plessis (University of Pretoria, South Africa), Yimin D. Zhang (Temple University, USA)
MIMO Radar Detection in Compound Gaussian Sea Clutter Using Joint Model and Sample Selection
Zhihang Wang, Zishu He and Qin He (University of Electronic Science and Technology of China, China)
This paper deals with the adaptive detection of moving target for multiple-input multiple-output (MIMO) radar in compound Gaussian sea clutter circumstances. A novel detector based on two-step generalized likelihood ratio test (GLRT) framework is developed by using model and sample selection (MSS) algorithm. Specifically, the type of detector is determined by the model selection (MS) approach according to the minimum of model fitting error using the Anderson-Darling (AD) test, and then the better secondary data are chosen from reference cells based on the samples selection (SS) criterion. Finally, the proposed detector on the basis of MSS algorithm is compared with traditional detectors using the real sea clutter data, i.e., IPIX data. The numerical results indicate that the novel MSS algorithm for detector exhibit more significant performances than the traditional detectors.
Parameter Optimization of Sparse Fourier Transform for Radar Target Detection
Hongchi Zhang and Tao Shan (Beijing Institute of Technology, China); Shengheng Liu (Southeast University & Purple Mountain Laboratories, China); Ran Tao (Beijing Institute of Technology, China)
The sparse Fourier transform (SFT) can dramatically accelerate the spectral analyses by leveraging the inherit sparsity in radar echoes. However, a satisfactory accuracy-complexity trade-off commonly requires sophisticated empirical parameter tuning. In this context, this work attempts to enhance SFT by optimizing the parameter selection mechanism. We first derive closed-form expressions of two performance metrics with respect to the detection and false-alarm rates. On top of this, a parameter optimization algorithm is designed. The proposed scheme is able to automatically arrive at a optimized parameter settings considering the a priori knowledge and the performance requirements, which is confirmed by numerical simulations.
Modified Iterative Adaptive Approach Based on Range-Doppler Matched Filter Outputs
Biao Zhang, Jing Tian and Siliang Wu (Beijing Institute of Technology, China)
Recently, iterative adaptive approach (IAA) has been adopted to obtain radar range-Doppler imaging with improved resolution and lower sidelobe levels, compared with conventional imaging methods. In this paper, we present a modified iterative adaptive approach based on range-Doppler matched filter outputs (MF-MIAA) within a small processing window. Since the covariance matrices of the range-Doppler cells within a certain processing window are the same with each other, the proposed method is able to reduce the computational cost by at least two orders of magnitude compared with original IAA and matched filter based iterative adaptive approach (MF-IAA), which is demonstrated by numerical simulations together with theoretical analysis.
Maritime Target Detection Using Frequency Estimation
Brian Ng (University of Adelaide, Australia); Luke Rosenberg (Defence, Science and Technology Group & University of Adelaide, Australia)
Coherent detection schemes have been shown to offer performance benefits over non-coherent techniques in the maritime radar context. To further improve performance, the use of sparse representations with judiciously selected dictionaries has been proposed as a method to separate target returns from surrounding sea clutter. This paper exploits a robust frequency estimation scheme to effectively separate the target from clutter and noise. The approach offers two advantages: (1) it does not require a dictionary to be designed to support the scheme and (2) the scheme is capable of estimating multiple frequencies, thus suitable for targets with multiple Doppler components. In this work, we use both real and simulated sea clutter with synthetic targets for evaluation. It is found that the proposed technique can improve performance over traditional coherent detection for both point targets and more complex targets with multiple frequency components.
Clutter-Ridge Matched SR-STAP Technique for Non-stationary Clutter Suppression
Hongda Ye (University of Electronic Science and Technology of China, China); Zhongyu Li (University of Electronic Science and Technology of China (UESTC), China); Zhutian Liu (University of Electronic Science and Technology of China, China); Junjie Wu and Haiguang Yang (University of Electronic Science and Technology of China (UESTC), China); Jianyu Yang (School of Electronic Engineering, China)
Non-stationary clutter suppression is an attractive and challenging problem in space-time adaptive processing(STAP) algorithms. Although sparse recovery(SR) STAP can effectively alleviate strong clutter non-stationary characteristic with a few number of samples, there are still difficulties where the clutter components are not located on the discrete space-time grid points of the dictionary, which results in significant performance loss. To overcome the off-grid problem, a clutter-ridge matched SR-STAP(CRM-SR-STAP) is proposed. In the proposed algorithm, clutter ridge is calculated at first according to prior knowledge including radar system parameters and modern inertial navigation system information. Then, the clutter-ridge matched dictionary is reconstructed to calibrate the uniformly discrete dictionary. Finally, the non-stationary clutter is suppressed by SR-STAP with this dictionary and the off-grid problem has been well-overcome. The proposed algorithm can effectively match the nonlinear characteristics of clutter under the non-side-looking array scenario, and the signal-to-interference-plus-noise ratio(SINR) loss also improves significantly. Simulations are given to verify the effectiveness of the proposed method.
Enhanced Graph-Based Detection for Moving Targets in Sea Clutter
Wenjing Zhao and Minglu Jin (Dalian University of Technology, China)
Signal processing on graphs offers a potential framework for signal features extraction by transforming the received signal into graph. Graph based information representation framework is explored and applied to radar target detection problem. The existing algorithm using graphs considers the case of long pulses and needs to perform eigenvalue decomposition, which requires high implementation complexity in practical scenarios. To render this algorithm practical, for the case of short pulses, this paper continues to explore the framework of signal processing on graphs, and designs an enhanced detection algorithm by using the ratio of maximum and minimum connectivity of graphs among the cell under test and reference cells as test statistic. Simulation results show that the proposed algorithm with a low computational complexity achieves superior detection performance than the existing algorithm based on graphs.
A Single Dataset Joint Domain Localized Algorithm for Clutter Suppression in Shipborne High Frequency Surface Wave Radar
Liang Guo, Xin Zhang, Yang Qiang, Lin Wang, Jiazhi Zhang and Weibo Deng (Harbin Institute of Technology, China)
Due to the motion of the platform, targets with low velocity like ships will be submerged in the spreading spectrum of the first-order sea clutter in shipborne high frequency surface wave radar (HFSWR). For shipborne HFSWR, the statistics characteristic of the first-order sea clutter vary significantly in range dimension and the qualified training samples are limited, the performance of the generally used clutter suppression method space-time adaptive processing (STAP) degrades in this scenario. To deal with this problem, a single dataset algorithm using only one range cell to estimate the clutter covariance matrix (CCM) is proposed. The CCM is estimated from other angle-Doppler regions in the same range bin based on the prior knowledge of the spread sea clutter. The proposed algorithm has a good performance tested by real data.
Coherent Fusion of Polarization Diversity Channels with Phase-Locked Loop for Target Detection
Yiheng Guo (Xidian University, China); Shenghua Zhou (National Laboratory of Radar Signal Processing, Xidian University, China); Aoya Wang (Xidian University, China); Hongwei Liu (National Laboratory of Radar Signal Processing, China); Yuehong Zhang (PLA Air Force Xi'an Flight Academy, China); Yunhe Cao (Xidian University Xi'an China, China); Hongtao Su (National Laboratory of Radar Signal Processing, China)
A full polarization radar generally has four polarization channels, whose signals are non-coherently accumulated in general. In this paper, a coherent fusion rule is presented for a full polarization diversity radar. It works in the target tracking mode and uses received signals to estimate and then remove the initial phases of target returns with the phase-locked loop, such that non-coherent target returns can be coherently combined. Simulation results indicate that this method can outperform the non-coherent one significantly in the target tracking mode.
Clutter Suppression and Moving Target Indication Using Deramp-STAP with Radon Transform for Multichannel HRWS SAR System
Tingting Zhang, Ke Tan, Xingyu Lu, Weimin Su, Hong Gu and Chenchen Wang (Nanjing University of Science and Technology, China)
Deramp space-time adaptive processing (Deramp-STAP) method can effectively reduce the number of channels used for azimuth ambiguity resolving in multichannel HRWS SAR system. However, traditional Deramp-STAP needs to be conducted in all the ambiguity interval of every frequency bin which results in a huge amount of calculation. Besides, it fails to correct the linear range migration (LRM) which deteriorates the result of imaging. In this paper, a novel Deramp-STAP method by using Radon Transform (RT) is proposed. Firstly, the STAP is operated when the searching ambiguity number is zero to suppress the clutter and extract the range information of the moving targets. After that, the equivalent relative speed of the moving target is estimated through RT and the frequency range of the moving target in azimuth Chirp Fourier Transform (CFT) frequency domain can be located. Also, LRM can be corrected due to the estimated equivalent relative speed. Then, the traditional Deramp-STAP is be performed in the narrowed frequency region for every searching ambiguity number and then the moving target can be indicated accurately. The effectiveness of the proposed method is validated by the simulation experiments.
A Novel Method to Suppress Short-Range Clutter in Airborne Radar
Fengde Jia, Ping Zhao, Lei Zhang and Menglin Zhai (Donghua University, China)
This paper primarily addresses the short-range clutter (SRC) suppression problem in non-side-looking airborne radar systems. For a high pulse repetition frequency, the generation of range-ambiguous clutter in the short range will affect the clutter covariance matrix (CCM) estimation accuracy. In addition, the CCM estimation requires a large number of range cells, and the elevation angle of the SRC changes rapidly with the range. Unlike the former method to design a series of pre-space-time adaptive processing (STAP) elevation filters (EFs), in which each EF sets a single null at the elevation angle of the short-range cell, a new method is proposed to design the EF only once by forming a certain null width to suppress the SRC. The modified EF design is based on average sidelobe level (SLL) control and constructed as a convex quadratic program, which can be solved using the CVX toolbox. The simulation results demonstrate that the proposed algorithm can effectively eliminate SRC while reducing the system implementation complexity.
Optimal Mismatched Filter Design for Radar Using Bounded Real Lemma
Tuomas Aittomäki and Visa Koivunen (Aalto University, Finland)
Mismatched filters can be used in remote sensing, particularly in radars, for sidelobe and interference attenuation. By correlating the received signal with not a copy of the transmitted waveform but a slightly modified signal, it is possible to adjust the sidelobes and interference level at the cost of reduced SNR. We show that the Bounded Real Lemma can be used to find the filter coefficients directly as a solution to a convex problem without need for any approximation or relaxation. Thus, contrary to the previous methods, the globally optimal mismatched filter can be found efficiently even when the Doppler shift of the received signal is significant, such as in the case of slow-time coded radar pulses.
Region Based Single-Stage Interference Mitigation and Target Detection
Anand Dubey (Friedrich-Alexander University Erlangen-Nürnberg, Germany); Jonas Fuchs (Friedrich-Alexander University Erlangen-Nuremberg, Germany); Venkat Madhavan (Friedrich-Alexander University, Germany); Maximilian Lübke (Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany); Robert Weigel (Friedrich-Alexander Universität Erlangen-Nürnberg, Germany); Fabian Lurz (Hamburg University of Technology, Germany)
The inherent smaller radar cross sections of vulnerable road users resulting in smaller signal-to-noise-ratios make an accurate detection of them somewhat challenging. Mutual radar interference in typical automotive scenarios further imposes the difficulty of a target detection by additionally raising the noise floor. The traditional signal processing pipeline consists of multiple but separate stages for interference detection, mitigation and target detection. In this paper, a convolutional neural network based autoencoder architecture is used to perform a combined single-stage target detection while generalizing over different interference noise. The proposed approach achieves significant improvement over state-of-the-art methods while preserving the instance of each target and is able to identify them uniquely in case of a partial occlusion or overlapping of multiple targets.
Broadened Sea Clutter Suppression Method for Shipborne HF Hybrid Sky-Surface Wave Radar
Xibo Zhou, Wei Yinsheng and Yong-Tan Liu (Harbin Institute of Technology, China)
In shipborne high frequency (HF) hybrid sky-surface wave radar, the sea clutter is the important component of the echoes and the major clutter background for vessel target detection. However, due to the complex radio paths, the sea clutter spectra spread severely and exhibit a complex signature. In this paper, a space-time adaptive processing (STAP) based on the Gabor wavelet filter to extract the sea clutter space-time distribution is proposed to suppress the broadened sea clutter. The proposed method involves three steps. First, the two-dimensional Gabor wavelet transform is applied to the Angle-Doppler(AD) spectrum to extract the sea clutter component by using the adaptive threshold method. Second, the noise region around the sea clutter is eliminated by using the template convolution and connected component labeling method. Third, the clutter covariance matrix is constructed based on the segmentation area of the sea clutter. According to the experimental results, the proposed method can obtain a more accurate estimation of the broadened sea clutter space-time distribution and achieve better suppression performance.

TM-P6: Space-based/Airborne radars

Chairs: Marta Bucciarelli (SYMPAS S. r. l. & University of Rome, "La Sapienza", Italy), Laura Candela (Italian Space Agency, Italy)
A Quad-Pol SAR Imaging Mode with Sound Azimuth Ambiguity
Yanyan Zhang (University of Chinese Academy of Sciences & Chinese Academy of Sciences, China); Yongwei Zhang (Institute of Electronics, Chinese Academy of Sciences & University of the Chinese Academy of Sciences, China); Yunkai Deng (Institute of Electronics, Chinese Academy of Sciences, China); Sheng Chang (University of Chinese Academy of Sciences, China); Dacheng Liu (Aerospace Information Research Institute, Chinese Academy of Sciences, China); Robert Yu Wang (Institute of Electronics, Chinese Academy of Sciences, China)
The azimuth ambiguity of the cross-polarized (cross-pol) echoes of hybrid quadrature-polarimetric synthetic aperture radar (quad-pol SAR) mainly comes from its co-polarized (co-pol) echo signals in the odd ambiguity regions. Here to improve the azimuth ambiguity of hybrid quad-pol SAR, a SAR imaging mode that alternately transmits the orthogonal elliptically polarized (OEP) waves to weight these co-pol echoes is detailed. Moreover, the boost of SAR azimuth ambiguity is described by the azimuth ambiguity-to-signal ratio (AASR) and evaluated by the L-band system parameters of spaceborne SAR. In addition, the chirp scaling algorithm (CSA) is adopted to execute the scene simulation experiments. And the results demonstrate that the presented SAR imaging mode can suppress the azimuth ambiguity of hybrid quad-pol SAR and verify the theoretical analysis.
Space-borne Sub-THz ISAR System for Objects with Translational Motion
Emidio Marchetti, Andrew Stove and Mikhail Cherniakov (University of Birmingham, United Kingdom (Great Britain)); Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain)); David Blacknell (DSTL, United Kingdom (Great Britain))
The exploitation of sub-terahertz technology (200 GHz -700 GHz) is proposed for imaging and the recognition of an object's state from a space-based sensor using inverse synthetic aperture radar, benefiting from wide bandwidth and enhanced sensitivity to surface texture of signals in this band. An initial technical analysis of the system is presented with the calculation of expected resolutions and of the power budget. Initial experimental results from measurements in a scaled environment are presented.
NewSpace SAR: Disruptive Concepts for Cost-Effective Earth Observation Missions
Michelangelo Villano, Nertjana Ustalli and Luca Dell'Amore (German Aerospace Center (DLR), Germany); Se-Yeon Jeon and Gerhard Krieger (DLR, Germany); Alberto Moreira (German Aerospace Center (DLR), Germany); Maxwell Nogueira Peixoto (Aeronautics Institute of Technology (ITA), Brazil); Jan Krecke (The University of Auckland, New Zealand)
Synthetic aperture radar (SAR) is a key remote sensing technique for Earth observation. Future SAR missions will deliver weekly high-resolution images of our planet, thereby allowing quantification of several essential climate variables. While this is a huge step forward compared to current systems, some applications require even more frequent temporal sampling or simultaneous acquisitions from slightly different observation angles. NewSpace SAR denotes all groundbreaking solutions that enable frequent and enhanced SAR imaging at affordable costs. Besides the technological developments, e.g., mass-produced platforms for constellations of SAR satellites, application-driven SAR system design approaches play a fundamental role. Moreover, disruptive concepts based on formations of alternately-transmitting satellites, waveform or phase encoding and multi-focus-post processing, allow relaxing the design constraints, reducing complexity, size, and cost of the SAR instrument, and still retrieving the desired information from SAR data. These solutions will help spreading the on-going NewSpace revolution to SAR remote sensing and posing the basis for future Earth observation missions that will yield remarkable societal benefits.
Effect of Wind Speed on Internal Wave Imaging of Multi-Polarimetric Spaceborne SAR
Zhang Yun, Tian Xiao, Lupeng Guo and Hongbo Li (Harbin Institute of Technology, China)
Kelvin, turbulence and internal waves are the three most common wake waves. Internal waves(IWs) are most easily detected because of their long propagation time and wide range. In this paper, the IWs model generated by moving objects and scattering theory are investigated to simulate IWs in SAR images. Since the wind speed acting on the wind-induced sea surface IWs model and the polarization mode acting on the scattering model will affect the SAR imaging of IWs, combining both, we propose a new method to analyze the effect of wind speed on IWs imaging of polarimetric SAR. In the proposed method, the effects of wind speed on individual polarization channels and multi-polarization features are discussed respectively. The two indexes of modulation depth and mean normalized difference are used to evaluate the imaging quality and imaging stability to obtain the optimal polarization features, which are beneficial to improve the accuracy of subsequent IW detection and parameters inversion of submerged object over the conventional analysis methods that only consider the individual polarization channels and imaging effects.
WIVERN: An ESA Earth Explorer Concept to Map Global In-Cloud Winds Precipitation and Cloud Properties
Anthony Illingworth (University of Reading, United Kingdom (Great Britain)); Alessandro Battaglia (Politecnico of Turin, Italy & University of Leicester, United Kingdom (Great Britain)); Julien Delanoe (Institut Pierre Simon Laplace, France)
The main objective of the proposed WIVERN mission is to provide global line-of-sight in-cloud winds in real time that can be assimilated into numerical weather prediction (NWP) models to improve weather forecasts. This will be achieved by a conically scanning dual polarisation Doppler 94 GHz radar with an 800 km wide ground track in a sun-synchronous polar orbit to provide daily visits poleward of 50°. According to the World Meteorological Organization (WMO), wind-storms are by far the largest contributor to economic losses caused by weather related hazards, resulting in approximately 500 billion USD (adjusted to 2011) of global damage over the last decade. A unique advantage of WIVERN is its ability to measure winds within active weather systems that are filled with thick cloud where there are currently very few wind observations, especially in tropical cyclones and hurricanes. These in-clouds winds will complement the predominantly clear air winds from the AEOLUS wind lidar launched in August 2018 which have been shown to have a major impact in reducing forecast errors. A subsidiary objective of WIVERN is to provide high resolution reflectivity profiles of rain, snow and ice water content to validate and improve parameterisation schemes in NWP and climate models.
Bayesian Azimuth Super-resolution of Sea-surface Target in Forward-looking Imaging
Yao Kang (University of Electronic Science and Technology, China); Yin Zhang, Deqing Mao, Xingyu Tuo, Yulin Huang and Yongchao Zhang (University of Electronic Science and Technology of China, China)
Aiming at the problem that real-aperture scanning radar in forward-looking imaging has low azimuth angular resolution for sea-surface target, this paper presents an angular super-resolution method based on the maximum a posteriori (MAP) criterion. Firstly, sea-surface clutter can be well fitted with Weibull distribution, so this paper derives a Weibull-based maximum likelihood (ML) estimation method based on Newton-Raphson iteration to effectively improve the azimuthal resolution. However, the Weibull-based ML method has limited robustness of noise suppression and randomly converges to the local optimal solution under low signal to clutter ratio. Therefore, this paper adds a sparse distribution as the prior distribution of sea-surface target, which can be regarded as a constraint term so that proposed MAP estimation method can not only obtain the better property of noise suppression, but also well converge to the global optimum. Finally, the simulation results are given to verify the performance of proposed method.
Context Semantic Perception Based on Superpixel Segmentation for Inshore Ship Detection in SAR Image
Rufei Wang and Fanyun Xu (University of Electronic Science and Technology of China, China); Jifang Pei (University of Electronic Science and Techonology of China, China); Zhang Qian (UESTC, China); Yulin Huang and Yin Zhang (University of Electronic Science and Technology of China, China); Jianyu Yang (School of Electronic Engineering, China)
The inshore ship detection of SAR image is a challenging task, due to inherent speckle noise, complex land background interference, identifiable sea clutter and few available shape feature. In this paper, we propose a context semantic perception method based on superpixel segmentation for inshore ship detection in SAR image. Firstly, the sea-land segmentation is used to quickly locate candidate regions, the ocean regions and buffer regions on both sides of the coastline are reserved, after that, the maximum stable extremum region (MSER) method is used for pre-screening. Then, the background slice of each candidate target is performed superpixel segmentation to extract contextual semantic information, to identify the inshore ships and false alarms. The Sentinel-1 spaceborne SAR images covering a coastal region are used to verify the proposed method. The experimental results show that the proposed method can realize accurate inshore ship detection under the complex sea-land mixed background.
Imaging Region Bound of Scanning Radar Angular Super-resolution on Motion Platform
Deqing Mao and Yongchao Zhang (University of Electronic Science and Technology of China, China); Junyu Zhu (UESTC, China); Yin Zhang and Yulin Huang (University of Electronic Science and Technology of China, China); Jianyu Yang (School of Electronic Engineering, China)
Airborne radar can observe a wide-region image with coarse resolution using a scanning antenna. Because the aperture size of the antenna on a motion platform is limited, super-resolution imaging methods based on different echo models are presented to improve its angular resolution. In this paper, the applicable scopes of the super-resolution imaging models are compared quantitatively. First, the traditional models of scanning radar, such as the approximate convolution model (ACM) and the complex phase-convolution model (CPM), are introduced, and we analyze the echo models in spatial frequency domain. Then, according to the comparison of the spatial bandwidth of the antenna pattern and that of the platform motion, the strict region bound and the relax region bound for scanning radar are analyzed. At last, the numerical region bounds are deduced to analyze the influencing factors. The bounds are significant for the working mode design and imaging method selection for a scanning radar. Simulations verify the analyzed region bound.
Forward Looking Airborne Radar for Landing Aid
Konstantin Alexandrovich Lukin (IRE NASU National Academy of Sciences of Ukraine, Ukraine); Joao Moreira (Embrier, Brasil); Sergiy Lukin (O. Ya. Usikov Institute for Radiophysics and Electronics, National Academy of Sciences of Ukraine, Ukraine & Università degli Studi di Napoli Parthenope, Italy)
A forward-looking 36 GHz radar for aircrafts and helicopters is presented. It uses the rotating synthetic aperture for obtaining the azimuth resolution and independent radars with different antenna elevation angles for obtaining the elevation resolution. Main radar functionalities are the guidance for landing phase including the three-dimensional view, strike avoidance, moving target indication and detection of stationary air and ground obstacles. This article describes the radar principle, its modeling and an implementation example for a given operational requirement.
Calibration and Weather Observation of a Dual-polarized Phased Array Line Replaceable Unit Radar Demonstrator
Pei-Sang Tsai (National Center for Atmospheric Research & NCAR, USA); Rodrigo Lebron (University of Oklahoma, USA); Jonathan Emmett, Adam Karboski and Christopher Burghart (National Center for Atmospheric Research, USA); Jorge Salazar (University of Oklahoma, USA); Scott Ellis, James Ranson and Eric Loew (National Center for Atmospheric Research, USA)
This paper provides an overall, detailed description of a C-band, 64-element Line Replaceable Unit (LRU) radar system. The LRU radar system was developed as a technology demonstrator to explore various phased array related technologies and investigate the technical requirements for the next-generation airborne phased array radar (APAR). The demonstrator served as a research platform through various calibration related topics, beamforming, and expansion to an end-to-end radar system.
Study of SAR Internal Wave Imaging Parameters Based on Experimental Data
Lupeng Guo, Bin Zhao, Zhang Yun and Tian Xiao (Harbin Institute of Technology, China)
Many scholars can invert internal waves (IWS) parameters through IWS SAR images by empirical mode decomposition and other methods. The inversion of depth, velocity and other parameters of underwater moving objects through SAR IWS images has become the target of current research. In this paper, by comparing the errors of half-open angle and wavelength between simulated and measured data under the same working conditions, we hope to lay a foundation for fitting the measured model with the simulated model and inverting the parameter information of the underwater object. In order to ensure the accuracy of the image, this paper discusses the optimal parameters of SAR IWS imaging, and then compares the measured data with the simulated data under different working conditions, analyses the error between the two at half-opening angle and wavelength to prove the consistency between the measured image and the simulated image.

Tuesday, September 22 2:15 - 3:00 (Europe/Rome)

Opening ceremony

Tuesday, September 22 3:00 - 4:00 (Europe/Rome)

Inaugural Plenary Talk - Marco De Fazio (Leonardo Spa, Italy): Radars at Leonardo Company: a 70 years long heritage in land, naval, airborne, and spaceborne systems. Looking backward, moving forward.

Chair: Alfonso Farina (Leonardo Company Consultant, Italy)

Tuesday, September 22 4:00 - 4:20 (Europe/Rome)

Coffee Break

Tuesday, September 22 4:20 - 6:00 (Europe/Rome)

TA-L2: Multistatic radars

Room: Ch2
Chairs: Mark E Davis (Medavis Consulting, USA), Vishal Monga (Pennsylvania State University, USA)
4:20 Extended Target Detection and Localization in 802.11ad/y Radars
Emanuele Grossi (University of Cassino and Southern Lazio & Consorzio Nazionale Inter-universitario per le Telecomunicazioni (CNIT), Italy); Marco Lops (University of Naples Federico II & CNIT - Consorzio Universitario Nazionale per le Telecomunicazioni, Italy); Antonia Tulino (Bell Labs, USA & Università Federico II, Italy); Luca Venturino (Universita' degli Studi di Cassino e del Lazio Merdionale & Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), Italy)
In this work, we consider an opportunistic mmWave radar obtained by adding a radar receiver to a 802.11ad/y device and we study the problem of detecting and localizing multiple extended (along both the range and Doppler dimensions) targets back-scattering the signal emitted by the communication transmitter. We derive an iterative subspace-based detector/estimator which extracts the prospective targets one-by-one from the superimposed received echoes, after cleaning up the previously-detected signal components. A short-range low-mobility application is discussed to validate the proposed solution.
4:40 A New Method and a Case Study in Statistical Modeling of Bistatic Radar Cross Section
Michał Meller (Gdansk University of Technology & PIT-RADWAR S.A., Poland); Maciej Wielgo and Mateusz Malanowski (Warsaw University of Technology, Poland); Jonathan Pisane (John Cockerill Defense, Belgium); Sylvian Azarian (SDR Technologies SAS, France)
We propose new tools that allow one to perform statistical modeling of radar cross section (RCS). In our approach, we model fluctuations of RCS as a realization of nonstationary random process with a hidden time-varying state that governs its local properties. We describe how one can employ a recently proposed Bayesian tracker to fit the adopted model to an observed sequence of data, and explain how to validate the fitted model's goodness of fit. We apply the proposed approach to data recorded with the PaRaDe FM passive radar system. The results support our hypothesis that long integration times that are typically employed in FM passive radars result in smoother (less spiky) behavior of target RCS than predicted by the classical Swerling I and III models.
5:00 Occupancy Detection and People Counting Using WiFi Passive Radar
Chong Tang, Wenda Li, Shelly Vishwakarma, Kevin Chetty and Simon Julier (University College London, United Kingdom (Great Britain)); Karl Woodbridge (University College London (UCL), United Kingdom (Great Britain))
Occupancy detection and people counting technologies have important uses in many scenarios ranging from management of human resources, optimising energy use in intelligent buildings and improving public services in future smart cities. Wi-Fi based sensing approaches for these applications have attracted significant attention in recent years because of their ubiquitous nature, and ability to preserve the privacy of individuals being counted. In this paper, we present a Passive Wi-Fi Radar (PWR) technique for occupancy detection and people counting. Unlike systems which exploit the Wi-Fi Received Signal Strength (RSS) and Channel State Information (CSI), PWR systems can directly be applied in any environment covered by an existing WiFi local area network without special modifications to the Wi-Fi access point. Specifically, we apply Cross Ambiguity Function (CAF) processing to generate Doppler spectrograms, and employ a CLEAN algorithm to remove the direct signal interference. A Convolutional Neural Network (CNN) and sliding-window based feature selection scheme is then used for classification. Experimental results collected from a typical office environment have validated that the proposed PWR system is able to determine room occupancy with 99.5\% accuracy, and correctly predict (98.1\%) the number of people when using four test subjects in experimental measurements.
5:20 Optimal Receiver Placement in Staring Cooperative Radar Networks for Detection of Drones
Benjamin D Griffin (Cranfield Defence and Security & Aveillant, United Kingdom (Great Britain)); Alessio Balleri (Cranfield University, United Kingdom (Great Britain)); Christopher J. Baker and Mohammed Jahangir (University of Birmingham, United Kingdom (Great Britain))
Staring radars use a transmitting static wide-beam antenna and a directive digital array to form multiple simultaneous beams on receive. Because beams are static, the radar can employ long integration times that facilitate the detection of slow low-RCS targets, such as drones, which present a challenge to traditional air surveillance radar. Typical low altitude trajectories employed by drones often result in low-grazing angle multipath effects which are difficult to mitigate with a monostatic radar alone. The use of multiple spatially separated receivers cooperating with the staring transmitters in a multistatic network allows multi-perspective target acquisitions that can help mitigate multipath and ultimately enhance the detection of drones. This paper investigates how varying the network geometry affects the estimation performance of a targets position and velocity in a multipath free scenario. The optimal geometry is found by minimising the trace of the Cramér- Rao Lower Bound (CRLB) of the Maximum Likelihood (ML) estimates of range and Doppler using the Coordinate Descent (CD) algorithm. The network estimation accuracy performance is verified using Monte Carlo simulations and an ML Estimator on the target parameter estimates.
5:40 DVB-T Based Forward Scatter Radar for Small Target Surveillance
Marco Di Seglio and Fabiola Colone (Sapienza University of Rome, Italy)
This paper investigates the target detection capability of a Passive Forward Scatter Radar (PFSR) exploiting a DVB-T transmitter as illuminator of opportunity. Specifically, it is shown that conventional processing schemes adopted in PFSR might suffer from significant performance degradation when dealing with OFDM waveforms of opportunity compared to the case of FM radio broadcast. In fact, a non-negligible increase of the disturbance background is observed in the final spectrogram, which yields an undesired masking effect on weak target echoes. Therefore, we propose a signal-based processing technique to mitigate the observed effect and improve the target detection capability. The proposed approach exploits the possibility to recover a good copy of the transmitted signal based on its digital nature; this is then used to cancel the fluctuating component of the output arising from the adopted modulation of the waveform of opportunity. The effectiveness of the proposed approach is proved against simulated data.

TA-L4: Short range civilian radars

Room: Ch4
Chairs: Francesca Filippini (Sapienza University of Rome, Italy), Anthony Martone (US Army Research Laboratory, USA)
4:20 Vibration Disturbance Cancellation Method for Estimation of Target Displacement by CW Doppler Radar
Takahiro Kinoshita (Nippon Steel Corporation & Niigata University, Japan); Hiroyoshi Yamada (Niigata University, Japan)
Recently, deterioration of infrastructures has become a problem, therefore a way of detecting the signs of deterioration by monitoring the conditions of the structures has been needed. Measuring vibrations of structures by using millimeter-wave radar that can be used at night and even in the rain is one of the useful methods. However, sometimes the radar itself is subject to disturbances such as vibration depending on the installation conditions of the radar when measuring structures outdoors. Hence, it is necessary to remove these disturbances and extract only the vibration of the target for measurement. In this paper, we propose a method for obtaining the amplitude and frequency of a target by taking the cross-spectrum of the baseband signals obtained by two CW Doppler radars mounted on different positions. Results of experiments suggest that the vibration disturbance of the platform can be eliminated properly based on this method.
4:40 A Comparison of the Recursive and FFT-Based Reassignment Methods in Micro-Doppler Analysis
Karol Abratkiewicz and Piotr Samczynski (Warsaw University of Technology, Poland); Dominique Fourer (IBISC, University of Evry/Paris-Saclay, France)
A brief comparison of two time-frequency (TF) reassignment methods is provided in this paper. The main advantage of the considered approach is that it allows energy concentration of the signal distribution on the TF plane to be obtained. Both techniques use the short-time Fourier transform (STFT), however, they can be formulated and computed differently. The first classical method is based on the fast Fourier transform (FFT), while the second one uses a recursive filter bank which, in turn, can be more efficient due to a lower time delay and a reduced computational complexity. Thanks to the proposed methodology, a real-time computation of the spectrogram and the reassigned spectrogram can be obtained. Hence, the reassignment method allows an almost ideal localization of the micro-Doppler signature components in a TF distribution to be obtained. Both approaches are presented, investigated, and validated using real-life radar signals in the form of micro-Doppler signatures originating from different targets.
5:00 Motion Blur Suppression Accommodating to Fast Radar Imaging for Walk-Through Concealed Weapon Detection
Tatsuya Sumiya and Kazumine Ogura (NEC Corporation, Japan); Shingo Yamanouchi (IEEE, USA); Nagma Khan, Masayuki Ariyoshi and Toshiyuki Nomura (NEC Corporation, Japan)
This paper presents a motion blur suppression technique that accommodates to fast radar imaging for walk-through concealed weapon detection. The proposed technique enables to generate radar images of a walking person in high-quality and at a high frame rate. High-quality imaging is achieved by compensating for the motion blur of the moving target. To achieve the high frame rate, the proposed technique is designed while keeping in mind the condition required for fast Fourier transform, so that the fast imaging algorithm can be applied. The motion blur suppression effects and computational efficiency are demonstrated both in numerical simulations and in experimental verifications by our developed walk-through radar imaging system. It is confirmed that the proposed technique well suppresses motion blur of moving targets and generates radar images in good quality. The processing time in our system is less than 100 ms, which is fast enough to run at a video rate.
5:20 Derivative Target Line (DTL) for Continuous Human Activity Detection and Recognition
Ronny Gerhard Guendel (Delft University of Technology, The Netherlands); Francesco Fioranelli and Alexander Yarovoy (TU Delft, The Netherlands)
In this paper, we investigate the classification of Activities of Daily Living (ADL) by using a pulsed ultra-wideband radar. Specifically, we focus on contiguous activities that can be inseparable in time and share a common transition, such as walking and falling. The range-time data domain is deliberately exploited to determine transitions from translation activities to in-place activities and vice versa, using a simple, yet effective approach based on the proposed Derivative Target Line (DTL). The separation of different in-place activities is then addressed using an energy detector finding the onset and offset times. Furthermore, the possible ADL for classification are limited at any decision stage based on kinematic constraints of human movements. We show that such limitation of classes at any given time leads to a classification improvement over a classifier containing always all ADL classes.
5:40 Hand Gesture Recognition via Radar Sensors and Convolutional Neural Networks
Stefano Franceschini (University of Naples Parthenope, Italy); Michele Ambrosanio (Università di Napoli Parthenope, Italy); Sergio Vitale (Università degli Studi di Napoli Parthenope, Italy); Fabio Baselice (Università degi Studi di Napoli Parthenope, Italy); Angelo Gifuni (Universita di Napoli PArthenope, Italy); Giuseppe Grassini (University of Naples Parthenope, Italy); Vito Pascazio (Università di Napoli Parthenope, Italy)
In this communication, a low-cost radar-sensor-based apparatus for contactless hand gesture recognition via Doppler signature analysis is proposed. The raw reflected signal, after some pre-processing, is analysed via its time-frequency representation, known as spectrogram. This information is then exploited to train a convolutional neural network (CNN) to perform the classification step. The whole procedure was tested on an in-house experimental data set composed of four different hand gestures, showing good performance and reaching an accuracy of approximately 97%. Finally, the classification performance was tested also in a cluttered environment which includes the presence of a strong echo close to the target.

TA-SS2: Topics, trends and challenges in cognitive radars

Room: Ch1
Chair: Graeme E Smith (The Johns Hopkins Applied Physics Laboratory, USA)
4:20 The Role of Cognition in Radar Sensing
Antonio De Maio (University of Naples "Federico II", Italy); Alfonso Farina (Leonardo Company Consultant, Italy)
The paper is aimed at discussing the role of cognition in radar sensing providing a system level point of view. In this respect some fundamental issues are considered involving transmitter, receiver, signal and data processing units. Firstly, the definition of the cognitive radar concept is provided and its key aspects are pinpointed. Secondly, the role played by the perception-action cycle and the awareness gleaned by dynamic databases and/or sensing networks is emphasized. Hence the application of the cognitive paradigm for radar operation in spectrally crowded environments is discussed and an example with the use of real hardware is shown. Finally a glimpse on the role of cognition in agile tracking systems is given.
4:40 Practical Aspects of Cognitive Radar
Anthony Martone (US Army Research Laboratory, USA); Kelly Sherbondy (Army Research Laboratory, USA); Jacob Kovarskiy and Benjamin Kirk (The Pennsylvania State University, USA); Jonathan Owen and Brandon Ravenscroft (University of Kansas, USA); Austin Egbert, Adam Goad and Angelique Dockendorf (Baylor University, USA); Charles E Thornton and Michael Buehrer (Virginia Tech, USA); Ram M Narayanan (Pennsylvania State Universiy, USA); Shannon D Blunt (University of Kansas, USA); Charles Baylis (Baylor University, USA)
In this paper we examine some of the practical aspects of implementing cognitive radar (CR) techniques onto software defined radar (SDRadar) platforms. These aspects include: 1) the response time (RT) of algorithms and components to determine latency bottlenecks, 2) autonomous regulation of the perception-action cycle (PAC) to determine "how fast the CR can interact with the environment" as well as "how fast the CR should interact with the environment," and 3) regulation of the cognition level to understand how to select a particular CR technique appropriately for a given dynamically-changing environment. To provide concrete examples of these three implementation aspects for CR, we will focus on the specific application of target tracking in a congested spectral environment.
5:00 Modeling and Simulation of Cognitive Radars
Sandeep Gogineni (ISL, Inc., USA); Joseph R. Guerci (Information Systems Laboratories, Inc. USA, USA); Hoan Nguyen (Information Systems Laboratories, USA); Jamie Bergin (ISL, USA); Brian Watson (ISL, Inc., USA); Muralidhar Rangaswamy (AFRL, USA)
Cognitive radar has emerged as key enabling technology to meet the demands of ever increasingly complex and congested radio frequency (RF) operating environments. The generally non-stationary, heterogeneous, and time-varying nature of the modern RF environment all but precludes the use of traditional adaptive processing methods that require the existence of wide sense stationary (WSS) training data. In this paper, an advanced modeling & simulation (M&S) framework is presented that captures much of the real-world physics that gives rise to the aforementioned challenges such as heterogenous clutter, dense background targets, and intentional/unintentional radio frequency interference (RFI).
5:20 Implementing Perception-Action Cycles Using Stochastic Optimization
Alexander Charlish (Fraunhofer FKIE, Germany); Kristine Bell (Metron, Inc., USA); Chris Kreucher (Centauri)
Cognitive radar problems involve the selection of actions based on the uncertain knowledge of a system state that is partially observed through noisy measurements. This process of sequential decision making under uncertainty can be considered as a stochastic optimization problem. This paper explicitly makes the connection between cognitive radar and stochastic optimization by presenting a framework for describing cognitive radar problems in terms of stochastic optimization, thereby pointing to ways to employ stochastic optimization for designing perception-action cycles in a cognitive radar.
5:40 Neural Networks & Machine Learning in Cognitive Radar
Graeme E Smith (The Johns Hopkins Applied Physics Laboratory, USA); Sevgi Z Zubeyde Gurbuz (University of Alabama & TUBITAK Space Technologies Research Institute, Italy); Stefan Bruggenwirth (Fraunhofer FHR, Germany); Peter John-Baptiste (The Ohio State University, USA)
This paper reports on how neural networks and machine learning can support the development of cognitive radar systems. We discuss the aspects of cognition that can be supported by neural networks, review the recent literature on the use of neural networks for radar and review the significant challenges to implementation. The paper concludes with an example where a neural network, trained using reinforcement learning, generates radar waveforms containing a 26 dB notch in the power spectral density. The notch location is specified using a spectral mask that is the input to the neural network.

TA-SS3: Quantum radar: real world experiments and new theory

Room: Ch3
Chairs: Bhashyam Balaji (DRDC-Ottawa, Canada), Fred E Daum (Raytheon, USA)
4:20 Microwave Quantum Illumination with a Digital Phase-Conjugated Receiver
Shabir Barzanjeh (University of Calgary, Canada); Stefano Pirandola (University of York, United Kingdom (Great Britain)); David Vitali (University of Camerino, Italy); Johannes Fink (IST, Austria)
Quantum illumination is a sensing technique that employs entangled signal-idler beams to improve the detection efficiency of low-reflectivity objects in environments with large thermal noise. The advantage over classical strategies is evident at low signal brightness, a feature which could make the protocol an ideal prototype for non-invasive scanning or low-power short-range radar. Here we experimentally investigate the concept of quantum illumination at microwave frequencies, by generating entangled fields using a Josephson parametric converter which are then amplified to illuminate a room-temperature object at a distance of 1 meter. Starting from experimental data, we simulate the case of perfect idler photon number detection, which results in a quantum advantage compared to the relative classical benchmark. Our results highlight the opportunities and challenges on the way towards a first room-temperature application of microwave quantum circuits.
4:40 Practical Advantage in Microwave Quantum Illumination
Nizar Messaoudi (University of Waterloo & Keysight Technologies, Canada); A. m. Vadiraj (University of Waterloo, Canada); Jérôme Bourassa (Qubic, Canada); Bhashyam Balaji (Defence R&D Canada, Canada); Christopher M. Wilson (University of Waterloo & Institute for Quantum Computing, Canada)
Broadly speaking, in quantum illumination we can say that a proposed protocol has a "quantum advantage" if it outperforms all possible classical protocols. In the optical domain of LIDAR, this is the most useful metric as lasers can routinely produce nearly ideal classical states of light at room temperature (RT). This is not the case in the microwave domain of RADAR where the photon energy is much less than the 300K thermal energy, meaning that a real RT microwave source will always be contaminated by significant thermal noise. Thus, it is not clear if it is technologically possible to produce an ideal classical microwave signal at RT. It is therefore interesting to ask if a microwave quantum illumination protocol can have a "practical advantage" compared to the best technologically feasible RT microwave source. In this paper, we look to frame this question more precisely. As a concrete example, we present experimental results showing that, contrary to recent claims in the literature [1], an entangled microwave source amplified by a cryogenic HEMT amplifier fails to obtain any performance advantage over a simply constructed RT source and, in facts, performs significantly worse. We present a simple theory which explains the experimental results and which offers guidance on how a practical advantage might be achieved.
5:00 A Comparison Between Quantum and Classical Noise Radar Sources
Robert Sven Jonsson (Chalmers University of Technology & Saab AB, Sweden); Roberto Di Candia (Aalto University, Finland); Martin Ankel (Chalmers University of Technology, Sweden); Anders Strom (Saab AB, Sweden); Goran L. Johansson (Chalmers University, Sweden)
We compare the performance of a quantum radar based on two-mode squeezed states with a classical radar system based on correlated thermal noise. With a constraint of equal number of photons N S transmitted to probe the environment, we find that the quantum setup exhibits an advantage with respect to its classical counterpart of √ 2 in the cross-mode correlations. Amplification of the signal and the idler is considered at different stages of the protocol, showing that no quantum advantage is achievable when a large-enough gain is applied, even when quantum-limited amplifiers are available. We also characterize the minimal type-II error probability decay, given a constraint on the type-I error probability, and find that the optimal decay rate of the type-II error probability in the quantum setup is ln(1 + 1/N S) larger than the optimal classical setup, in the N S ≪ 1 regime. In addition, we consider the Receiver Operating Characteristic (ROC) curves for the scenario when the idler and the received signal are measured separately, showing that no quantum advantage is present in this case. Our work characterizes the trade-off between quantum correlations and noise in quantum radar systems.
5:20 Simulation Study of a Detector Function for QTMS Radar and Noise Radar
David Luong (Carleton University, Canada); Bhashyam Balaji (DRDC-Ottawa, Canada); Sreeraman Rajan (Carleton University, Canada)
A detector function for a newly proposed quantum radar known as quantum two-mode squeezing (QTMS) radar is analyzed in this paper. The detector function proposed in this paper behaves like a normal distribution function when the number of radar samples integrated for the detection is over 500 samples. A closed form expression for the receiver operating characteristic (ROC) curve for QTMS radar is derived under the assumption that the radar integrates more than 500 samples to detect the presence or absence of a target. The detector function can also be applied to noise radar.
5:40 Multiple Input-Multiple Output Quantum Radar
Marco Frasca (MBDA Italia, Italy); Alfonso Farina (Leonardo Company Consultant, Italy)
We consider a quantum radar, using as sources for entangled photons devices with spontaneous parametric down-conversion, with a given number of channels. This conception is similar to a MIMO radar and we exploit the detection theory for this case to show a neat improvement in the receiver operating characteristics (ROC) curves with increasing number of emitting sources. This appears to go against the prevalent trend in some quantum radar model that show degrading performances at increasing number of photons.

TA-SS4: Distributed SAR systems and missions

Room: Ch5
Chairs: Gerhard Krieger (DLR, Germany), Alberto Moreira (German Aerospace Center (DLR), Germany)
4:20 An Iso-Frequency MIMO SAR Formation for Wide-Swath Imaging Interferometry and Tomography
Davide Giudici (Aresys srl, Italy); Pietro Guccione (Politecnico di Bari, Italy); Marco Manzoni and Andrea Monti-Guarnieri (Politecnico di Milano, Italy); Fabio Rocca (Politecnico di Milano (emeritus), Italy)
The paper proposes a close formation of Synthetic Aperture Radar (SAR) satellites that are simultaneously transmitting and receiving, in a Multiple Input Multiple Output (MIMO) configurations. The received signals can be jointly processed to form a single SAR image with a power gain proportional to the squared number of sensors, or to upsample a low PRF, enabling wide coverage, or finally to retrieve a 3D complex reflectivity, in vertical layers, by a tomographic MIMO configuration. Evaluation of performance and limitations has been carried out in the three cases by running full 2D simulations and focusing.
4:40 Concepts and Applications of Multi-statitc MirrorSAR Systems
Josef Hermann Martin Mittermayer (German Aerospace Center (DLR), Germany); Gerhard Krieger (DLR, Germany); Alberto Moreira (German Aerospace Center (DLR), Germany)
The paper describes the basic components of MirrorSAR and explains how bi- and multistatic SAR acquisitions are achieved by shifting only minimal functionality to the receive satellites. It shows that synchronization is not required or reduced in complexity by a MirrorLink. The MirrorLink forwards the ground reflected radar signal from the receiving satellites in a space transponder manner to the transmit satellite. Several options for the MirrorLink are discussed. In MirrorSAR, a number of satellites enable the acquisition of dual- or multi-baselines in a single-pass. The paper discusses several application examples where MirrorSAR eases the overall system complexity, the data acquisition and improves the SAR product quality. An interferometric SAR application example is discussed in more detail.
5:00 A Highly Flexible and Scalable S-band SwarmSAR from Very Simple Nodes
Lorenzo Iannini (Delft University of Technology, Italy); Paco Lopez Dekker and Peter Hoogeboom (Delft University of Technology, The Netherlands)
The paper introduces the principles and the technical elements supporting the so-called SwarmSAR concept, consisting in a close formation of simple nodes cooperating in a MIMO-like frame to boost their imaging flexibility and performance. The philosophy of the swarm consists in employing extremely basic but self-sufficient nodes, each one guaranteeing hence sufficient image quality even when used individually. The costs are hence diverted from the node to the formation launching and maintenance aspects. We promote in this paper the use of S-Band as a convenient frequency both for the single node and for the formation requirements. An outline of the envisioned cooperative illumination modes, including high resolution imaging and the interferometric modes, and a preliminary discussion on their expected performance and challenges is provided.
5:20 Timing and Design Issues in Formation Flying Distributed SAR
Alfredo Renga, Maria Daniela Graziano, Marco Grasso and Antonio Moccia (University of Naples Federico II, Italy)
The paper investigates timing aspects of the design and the operations of a formation flying distributed Synthetic Aperture Radar (SAR). The analyzed system considers a satellite cluster including one transmitter and N receivers flying in formation with a dominant along-track baseline. This realizes a sensor receiving N samples at azimuth-displaced positions for each transmitted pulse, enabling several applications. In fact, besides the possibility to implement single-pass multi-baseline interferometric SAR techniques, the bistatic raw data collected by each receiver can be combined to generate a single higher performance image, e.g. High-Resolution Wide-Swath (HRWS) imaging. The paper investigates the latter topic, proposing dedicated solutions to preserve the imaging performance against the effects of relative orbit dynamics and residual orbit control errors. In details, the factor N can be interpreted as the redundancy, or the maximum number of degrees of freedom of the distributed system when compared to a single-channel monostatic SAR. Such a redundancy can be exploited in different ways pending on the application of suitable signal reconstruction and it represents an important parameter affecting the system timing. In this contest, the manuscript proposes an approach to system timing, satisfying not only ambiguity-related issues but also requirements concerning signal parameters, like SNR and\or Peak-to-Side-Lobe Ratio (PSLR). Finally, the timing considerations are recast in the framework of the overall design of a formation flying SAR.
5:40 Polychromatic Time Domain Reconstruction Approach for Along-track Multistatic SAR Constellations with Varying PRI
Nida Sakar (German Aerospace Center, Oberpfaffernhofen, Germany); Marc Rodriguez-Cassola (DLR, Germany); Pau Prats and Alberto Moreira (German Aerospace Center (DLR), Germany)
Next-generation spaceborne Synthetic Aperture Radar (SAR) systems aim at very high resolution imaging over wide swaths in the order of hundreds of kilometers. One viable realization of such a concept is distributed SARs with along-track baselines operated under the Nyquist frequency. Undersampled received data stack needs to be reconstructed to obtain azimuth-ambiguity free data on a regular sampling grid. On the other hand, very high-resolution distributed SAR system requires very precise orbit control to avoid sampling instabilities in along-track and phase error due to the topographic sensitivity. Both time and Doppler domain reconstruction methods in the literature perform well in case of uniform and nonuniform sampling conditions. However, highly nonuniform sampling and coinciding sampling result in severe signal-to-noise ratio degradation or impossibility to recover the entire Doppler bandwidth. Depending on the frequency band, even a small cross-track baseline in the order of few meters may result in high phase errors over mountainous terrain. In this paper, we analysed the impact of the varying pulse repetition interval (PRI) on the azimuth sampling condition and proposed two time-domain polychromatic reconstruction methods suitable for multistatic SAR constellations with varying PRI.

Wednesday, September 23

Wednesday, September 23 9:00 - 10:40 (Europe/Rome)

WM-L2: Radar waveform design/optimization

Room: Ch2
Chairs: Augusto Aubry (Universita degli studi di Napoli, Italy), Guolong Cui (University of Electronic Science and Technology of China (UESTC), China)
9:00 Transceiver Design in Signal-Dependent Interference and Spectrally Dense Environments
Jing Yang (University of Electronic Science and Technology of China, China); Augusto Aubry (Universita degli studi di Napoli, Italy); Antonio De Maio (University of Naples "Federico II", Italy); Xianxiang Yu (University Of Electronic Science And Technology Of China, China); Guolong Cui (University of Electronic Science and Technology of China (UESTC), China)
This paper deals with the joint design of transmit waveform and receive filter improving radar Signal to Interference plus Noise Ratio (SINR) in signal-dependent interference and spectrally contested-congested environments. The injected interference energy on each common band is precisely controlled for spectral compatibility. Besides, the transmit signal complies with constant envelop and similarity constraints to bestow some attractive waveform characteristics. To handle the resulting transceiver design problem, an iterative optimization procedure with a polynomial computational complexity is developed leveraging the Coordinate Descent (CD) framework. Finally, numerical results are provided to highlight the effectiveness of the proposed joint transmitter and receiver design technique.
9:20 Radar Waveform Synthesis Using Generative Adversarial Networks
Vesa T Saarinen and Visa Koivunen (Aalto University, Finland)
In this paper we propose a machine learning approach based on generative adversarial networks (GAN) for synthesizing novel radar waveforms with a desirable Ambiguity Function (AF) shape and constant modulus property. There are only a limited number of code sequences of a certain length in many widely used radar code families, which may be a drawback in modern radar applications. Hence, there is a need to generate new waveforms for future MIMO, multifunction, and cognitive radars. In such systems multiple waveforms are launched simultaneously in order to deal with low observable targets or a large number of small targets. Additionally, the ability to generate new waveforms at will makes it more difficult for an adversary to recognize or detect that it is illuminated by a radar. A Wasserstein GAN (WGAN) structure is developed for complex-valued input data. The model is trained using Frank and Oppermann codes with good autocorrelation and cross-correlation properties. The synthesized novel waveforms have an almost identical AFs to those of the training data, as well as a low cross-correlation relative to the codes in the training set. Additionally, the constant modulus property facilitates the efficient use of amplifiers.
9:40 Simultaneous Unambiguous Range and Doppler Through Non-Uniform Sampling
Warren P du Plessis (University of Pretoria, South Africa)
Range and Doppler ambiguities cause a variety of problems in radar systems, to the point of causing blind zones that can prevent target detection in some circumstances. Avoiding ambiguities in radar systems is thus desirable, but current approaches to achieving this goal have limitations. An approach to simultaneously achieving unambiguous range and Doppler based on non-uniform bursts is presented. Measurements using a sonar in air in the audio frequency range confirm the potential of this approach to eliminate ambiguities.
10:00 Optimization of Pulse-agile Sequence for Suppressing Range Folded Clutter Based on Diverse NLFM Waveforms
Ying-hao Sun, Hua-yu Fan, Ruo-dan Zhuang, Li-xiang Ren, Er-ke Mao and Teng Long (Beijing Institute of Technology, China)
Waveform diversity, which can be used to resolve range ambiguity and suppress range folded clutter in pulse-Doppler radar system, has attracted an increasing amount of attention in recent years. Especially for high pulse repetition frequency (PRF), the range folded clutter severely impacts the detection of small targets with long range and low speed. Using diverse waveforms, pulses of the transmitted sequence of radar are agile. It is possible to construct corresponding matched filters (MFs) to suppress the range folded clutter. However, the performance greatly depends on the cross-correlation among pulses of the transmitted sequence. In this paper, diverse nonlinear frequency modulation (NLFM) waveforms, the spectra of which are based on different window functions, are designed to construct a signal library, and positive and negative frequency modulation rate are both considered. Then, a Min-Max algorithm is proposed for sequentially selecting NLFM waveforms from the signal library to generate an optimal transmitted sequence. Pulse-to-pulse random initial phases are also added to each pulse of the sequence to further improve the irrelevance among pulses. Moreover, to eliminate the range sidelobe modulation (RSM) effect in Doppler processing, a cyclic algorithm for designing joint mismatched filters (JMMFs) with finite impulse response (FIR) is provided. Simulations show that the optimized pulse-agile sequence based on NLFM waveforms yields satisfactory performance on the ambiguity function, and range folded clutter can be effectively suppressed.
10:20 Joint Optimization of the Spectral Coexistence of Radar and Communications System
Junhui Qian, Shuya Zhang and Mengchen Lu (Chongqing University, China); Fengchun Tian (College of Communication Engineering,Chongqing University, China)
In the present paper, we proposed and analyzed a novel algorithmic framework for the joint design of the spectral coexistence of radar and communications system. The objects of interest in system design are the radar transmit code and the communication system code-book, whereby we state the design problem as a constrained maximization of the communication mutual information (MI) subject to a number of constraints guaranteeing both the radar performance, through a similarity constraint with some standard waveform and the effective power of radar interference. An iterative procedure, whose convergence is analytically proved, based on iterative alternating optimization of the suitably designed subproblems, is proposed and analyzed. Finally, we assess the effectiveness of the proposed approaches by numerical results.

WM-L4: Classification and recognition of drones

Room: Ch4
Chairs: Alessio Balleri (Cranfield University, United Kingdom (Great Britain)), Daniel A Brooks (Thales - LIP6, France)
9:00 Machine and Deep Learning for Drone Radar Recognition by Micro-Doppler and Kinematic Criteria
Frederic Barbaresco (Thales Air Systems, France); Daniel A Brooks (Thales - LIP6, France); Claude Adnet (Thales Air Systems, France)
Illegal, malicious or dangerous uses of drones, require developing systems capable of detecting, tracking and recognizing them in a non-collaborative way, and with enough anticipation in order to assign adapted interception means to the threat. The reduced size of autonomous aircraft makes it difficult to be detected over long distances with sufficient awareness based on conventional techniques, and seems more suitable for observation by radar sensors. However, the radiofrequency detection of this kind of object poses other difficulties to be solved due to their slow speed characteristics which can cause confusion with other mobile echoes like land vehicles, birds and vegetation movements agitated by atmospheric turbulence. It is therefore necessary to design robust classification methods for these echoes to ensure their discrimination relative to criteria characterizing their movements (micro-movements of their moving parts and kinematic movements of their main body).
9:20 Deep Learning and Information Geometry for Drone Micro-Doppler Radar Classification
Daniel A Brooks (Thales - LIP6, France); Olivier Schwander (École Polytechnique, France); Frederic Barbaresco (Thales Air Systems, France); Jean-yves Schneider (Thales, France); Matthieu Cord (Sorbonne Universités, France)
In this work, we build dedicated learning models for micro-Doppler radar time series classification. We develop both deep temporal architectures based on time-frequency representations, and also directly study the signal's underlying statistical Gaussian process using Information Geometry on Riemannian manifolds by developing and improving symmetric positive definite (SPD) neural networks. We also propose the aggregation of all proposed models in a single, highly performing classification pipeline.
9:40 Examination of Drone Micro-Doppler and JEM/HERM Signatures
John M Markow (United Kingdom (Great Britain)); Alessio Balleri (Cranfield University, United Kingdom (Great Britain))
Radars tasked with monitoring small targets often must increase their integration times to generate a useable signal-to-noise ratio (SNR) for maintaining a viable track. These longer integation times can prevent micro-Doppler signature extraction and instead result in Doppler signatures consisting of spectral lines to the radar's higher-level processing. Whether the radar operates in the micro-Doppler or spectral line regime depends on both radar parameters (e.g. waveforms, wavelengths and integration times) as well as target parameters (e.g. rotor length, rotational frequency, target reflectivity and geometry). This paper presents the use of modeling, simulations and experimental data to refine the understanding of how a particular radar will observe a target Doppler signature in either of these regimes, highlighting the transition region between the two.
10:00 Parametric Investigation on Simulated Staring FMCW Radar for Anti-Drone Swarms
Joongsup Yun and David Anderson (University of Glasgow, United Kingdom (Great Britain)); Francesco Fioranelli (TU Delft, The Netherlands)
This paper presents parametric investigation results on a staring FMCW radar system which targets drone swarms. The parametric investigation has been carried out by using the RAPID-SIM which facilitates system-level analysis of drone swarms' radar signatures. This paper explains concepts of the simulator's each module and also covers two parametric investigation results which deal with quantitative performance criteria for the design of the anti-drone swarms radar system.

WM-SS5: Sparse array design techniques for radar applications

Room: Ch1
Chairs: Elias Aboutanios (University of New South Wales, Australia), Xiangrong Wang (Beihang University, China)
9:00 Learning Sparse Array Capon Beamformer Design Using Deep Learning Approach
Syed Ali Hamza and Moeness G. Amin (Villanova University, USA)
The paper considers receive beamforming for de- signing sparse array configurations. We consider to maximize signal-to-interference plus noise ratio (MaxSINR) for a desired point source in a narrowband interferening environment. The sparse array design methods can either be data driven or rely entirely on the prior knowledge of the interference DOAs and respective powers. In this paper, we propose a design which is essentially data-driven and conceived by training the Deep Neural Network (DNN). Towards this goal, the training scenarios are generated through enumeration to learn an effective representation that can perform well in a downstream task. The input to the DNN is the received correlation matrix and output is the corresponding sparse configuration with superior interference mitigation capability. The performance of the DNN is evaluated by the ability to learn the implementation of enumerated design. It is shown through design examples that DNN effectively learns the enumerated algorithm and, as such, paves the way for efficient real-time implementation.
9:20 Online Antenna Selection for Adaptive Beamforming in MIMO Radar
Hamed Nosrati (CSIRO, Australia); Elias Aboutanios (University of New South Wales, Australia)
Antenna selection offers a low-cost solution for adaptive beamforming in large-scale multiple-input multiple-output (MIMO) radar as it alleviates the high cost of digital signal processing (DSP), and radio frequency (RF) stages. However, solving the selection problem is computationally expensive and scenario dependent. In this paper, we propose an online method for antenna selection to harness the power of unsupervised learning for reducing the computational cost. We first formulate the problem as an online game in the context of online convex optimization (OCO), and employ a gradient-based method that makes a decision for each new realization of the signal environment to minimize the total loss after T steps. We study the performance of the proposed method through a numerical example and compare the results with that of the global solution obtained by exhaustive search and an approximated solution provided by offline maximization. We demonstrate that the proposed online method is able to offer a low complexity solution that tracks the global solution, converging to the offline approximation method but with significantly reduced computational complexity.
9:40 Cognitive Wideband Beamforming for Sparse Array
Qinghui Lu, Ruitao Liu and Lifang Feng (University of Electronic Science and Technology of China, China); Guolong Cui (University of Electronic Science and Technology of China (UESTC), China); Zhenghong Zhang and Lin Zhou (Southwest China Research Institute of Electronic Equipment, China)
This paper deals with the constrained design of the weighting coefficients to achieve a desired receiving beampattern for wideband sparse phased array. A novel optimization model that minimizes beampattern peak sidelobe level in a space-frequency region of interest (SFRI) along with mainlobe level and weight coefficients quantitative restrictions is proposed. An Alternating Direction Penalty Method (ADPM) formulation to tackle the resultant nonconvex problem is introduced. In each iteration, the original problem is executed via solving multiple tractable subproblems. The simulation results highlight that in comparison with the available approach, the proposed method realizes a deeper nulling in the given space-frequency region capable of rejecting the strong interference.
10:00 Covariance Matrix Reconstruction Using Parsimonious Measurements and Low-Sample Support
Aboulnasr Hassanien (Wright State University, USA); Moeness G. Amin (Villanova University, USA)
In this paper, we consider the problem of spacetime adaptive processing (STAP) weight vector design using parsimonious spatial measurements and low temporal sample support. The extreme case when a single space-time data snapshot is the only available data is considered. It is assumed that dense discrete clutter components are spread along the clutter ridge. We propose a method for clutter-plus-noise covariance matrix reconstruction in the absence of secondary data. A two stage approach is adopted where a coarse angle-Doppler map is created in the first stage while a fine map is obtained in the second stage. It is shown that the clutter components can be accurately localized in the final angle-Doppler map. The final map is used to construct the covariance matrix and design the full array STAP weight vector. We show that the performance of the proposed STAP processor using parsimonious measurements is comparable to the performance of STAP design using full-dimensional arrays. Simulations examples are used to validate the effectiveness of the proposed STAP design technique.
10:20 Deterministic Sparse Array Capon Beamformer Design for Mitigating Spatially-Closed Interferences
Xiangrong Wang (Beihang University, China); Maria S. Greco and Fulvio Gini (University of Pisa, Italy)
Sparse arrays have attracted increased attention due to their capability of striking the best compromise between performance and complexity. High spatial resolution associated with a large aperture makes sparse arrays most effective in combating spatially closed interferences, which are notoriously deteriorative. The configuration of sparse arrays achieving the maximum array gain generally varies with changing scenario, which is referred to as reconfigurable sparse arrays. On the contrary, deterministic sparse arrays have a fixed structure, thus more preferable in practical applications due to its hardware efficiency, although not optimal in terms of adaptive interference mitigation. Thus, it is preferred that deterministic sparse arrays could be designed for mitigating spatially closed interferences. In this work, we first eliminate the dependence of the array gain on array configurations utilizing the relationship between the spatial correlation coefficient (SCC) and array gain, in turn transforming the problem of reconfigurable arrays to that of deterministic arrays. We then propose a modified Alternative Direction Method of Multipliers (ADMM) algorithm to solve the sparse array design problem. We also compare the modified ADMM algorithm with our previously proposed Difference of Convex Sets (DCS) method, and results show that the modified ADMM is insensitive against initial search points and outperforms the previously proposed DCS method.

WM-SS6: Emerging technologies in automotive radars

Room: Ch3
Chairs: Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain)), Bhavani Shankar Mysore R (Interdisciplinary Centre for Security, Reliability and Trust & University of Luxembourg, Luxembourg)
9:00 STAP in Automotive MIMO Radar with Transmitter Scheduling
Guohua Wang and Siddhartha Siddhartha (Hertzwell Pte Ltd, Singapore); Kumar Vijay Mishra (United States Army Research Laboratory, USA)
Automotive radars often employ multiple-input multiple-output (MIMO) array to attain high angular resolution with few antenna elements. The diversity gain is generally achieved by time-division multiplexing (TDM) during the transmission of frequency-modulated continuous-wave (FMCW) signals. However, TDM mode leads to longer pulse repetition intervals and, therefore, inherently and severely limits the maximum unambiguous Doppler velocity that a radar is able to detect. In this paper, we address the Doppler ambiguity problem in TDM MIMO automotive radars through a space-time adaptive processing (STAP) approach. A direct application of STAP may lead to a high antenna sidelobe level that hampers the detection performance. We mitigate this through optimal transmitter scheduling. We formulate the problem as combinatorial optimization and solve it via reinforcement learning. Numerical and experimental results demonstrate the efficacy of our method when compared with conventional techniques.
9:20 Adaptive Integration Time in Automotive SAR
David Wright and Shahzad Gishkori (University of Edinburgh, United Kingdom (Great Britain)); Liam Y. Daniel (University of Birmingham, United Kingdom (Great Britain)); Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain)); Bernard Mulgrew (Institute for Digital Communications, The University of Edinburgh, United Kingdom (Great Britain))
In this paper, we propose an adaptive sub-aperture integration approach for wide-angle synthetic aperture radar (SAR) with emphasis on the automotive applications. Traditional SAR integration approaches use fixed-width sub-apertures. These approaches suffer from the fact that scatterers may have different persistence widths over the angular aperture and can result in spreading the response across multiple sub-apertures or, conversely, coherently integrate extra noise with the signal. Our proposed method employs change-point detection to identify the persistence widths of the scatterers and consequently increases the coherent integration gain. Experimental results on two datasets validate our proposed methodology.
9:40 Simulator Design for Interference Analysis in Complex Automotive Multi-User Traffic Scenarios
Lizette Lorraine Tovar Torres and Fabian Roos (Ulm University, Germany); Christian Waldschmidt (University of Ulm, Germany)
Since the number of radar sensors per vehicle has been increasing in order to improve the performance of Advanced Driver Assistance Systems, mutual interference has become one of the most important challenges for near future automotive radar systems. Depending on the sensor parameters and the diverse properties of the traffic scenario, the interference effects can significantly vary. The effort to measure and appraise the interference in real traffic scenarios leads to the importance of designing a tool suited to simulate and analyze the interference. This paper presents a fast simulator capable of modeling, estimating, and evaluating the effects of interference in complex multi-user traffic scenarios. A statistical analysis of the interference in a proposed scenario is performed in order to validate the presented simulator.
10:00 Effect of Volume Distributed Clutter in PSK Modulated Automotive Radar
Marco Bersaglia (Radar and Surveillance Systems National Lab - CNIT, Italy); Fatemeh Norouzian and Edward Hoare (University of Birmingham, United Kingdom (Great Britain)); Marco Martorella (University of Pisa, Italy); Mikhail Cherniakov (University of Birmingham, United Kingdom (Great Britain))
Due to rapid progress in silicon technology, a number of companies have developed automotive radars which use Phase Shift Keying modulated signals. The effect of Volume Distributed Clutter (VDC), e.g. rain, hail, etc. on the radar performance is considered in this paper. It is shown that the sidelobe levels of the signal Ambiguity Function dramatically affects the radar performance, and modulation codes should be selected taking into account the presence of VDC. The methodology of the system performance evaluation is presented and accompanied with practical examples.

WM-SS7: Multitemporal SAR image processing and analysis

Room: Ch5
Chairs: Sebastiano Serpico (University of Genoa, Italy), Emmanuel Trouvé (Univ-smb, France)
9:00 Urban Change Pattern Exploration Using Fine-Resolution SAR of Ascending and Descending Orbits
Meiqin Che (University of Pavia, Italy); Paolo Gamba (Università degli Studi di Pavia, Italy)
In the last few decades, worldwide urbanization has promoted the emergence of megacities, megalopolises, urban clusters or large urban aggregations. It brings greater global challenges and climatic, environmental, socio-economical issues. Thanks to advanced remote sensing technologies, the necessary and urgent need to monitor urban changes can be performed by using multi-temporal and spatially fine-resolution data of increasing availability.
9:20 Change Detection in Single- And Multi-Look Polarimetric SAR Data
Allan A Nielsen, Paul J Connetable, Knut Conradsen and Henning Skriver (Technical University of Denmark, Denmark); Ernst Krogager (Danish Defence Acquisition and Logistics Organization (DALO) & LU-VV06, Denmark)
We describe the use of P-band data from DLR's F-SAR system to change detection between three time points. Change detection on data in the covariance matrix representation is performed by means of test statistics in the complex Wishart distribution on both multi- and single-look data. We successfully detect objects including humans buried under snow in Disko Island, Greenland.
9:40 Unsupervised Heterogeneous Change Detection in Radar Images by Cross-Domain Affinity Matching
Stian Normann Anfinsen, Luigi Tommaso Luppino and Mads Hansen (UiT The Arctic University of Norway, Norway); Gabriele Moser (Università degli Studi di Genova, Italy); Sebastiano Serpico (University of Genoa, Italy)
A new methodology for unsupervised heterogeneous change detection has recently been proposed, which combines deep neural networks for domain alignment and image-to-image regression with a comparison of domain-specific pixel affinities to reveal structural changes. In this paper we explain the underlying cross-domain dissimilarity measure and give an example of how it can be integrated in a deep neural network architecture and used to strengthen approaches based on image transformations. We demonstrate that the method is viable for bitemporal change detection with synthetic aperture radar (SAR) images combined with optical images, as well as with combinations of multifrequency and multipolarisation SAR images, but also provide a critical assessment of weaknesses and remaining challenges.
10:00 Analysis of Dense Coregistration Methods Applied to Optical and SAR Time-Series for Ice Flow Estimations
Laurane Charrier (ONERA/LISTIC, France); Pierre Godet (ONERA, France); Clément Rambour (ONERA/CNAM, France); Flora Weissgerber, Simon Erdmann and Elise Colin Koeniguer (ONERA, France)
This article considers the application of two dense coregistration algorithms to the estimation of ice flow. These algorithms estimate displacements at each pixel of the image and can be applied to pairs of radar, optical and radar/optical images. This flexibility combined with the dense estimation should improve both spatial and temporal resolutions of glacier displacement maps. Several tests are carried out on Sentinel-1 and Sentinel-2 images on Totten glacier in Antarctica. We assess the reliability of the considered algorithms by applying them to real and emulated pairs of images based on displacement fields previously estimated in the literature.
10:20 A Geometrical Wavelet Framework for the Time-Series Analysis of Full-Polarimetric Features
Davide Pirrone (LISTIC, University Savoie Mont Blanc, Annecy, France); Abdourrahmane M. Atto (LISTIC, Université Savoie Mont Blanc, France); Emmanuel Trouve (University Savoie Mont Blanc, France)
Polarimetric SAR (PolSAR) image time series have been employed for the analysis of temporal patterns of natural features in terms of the extended polarimetric scattering properties. However, the time series provide a rich scattering information that can be used for tracking and analyzing the evolution of targets, individuating smooth and/or abrupt changes. In this work we propose a wavelet framework that exploits the information from polarimetric features and analyze them to both mitigate the speckle effect on the multi-temporal information and improve the targets homogeneity using the multi-temporal information. The framework combines the powerful description from the main polarimetric decomposition features and the temporal analysis using geometrical wavelet transform. The analysis is applied on a multi-temporal polarimetric dataset of Radarsat-2 images acquired over the Argentiere glacier site.

Wednesday, September 23 10:40 - 11:00 (Europe/Rome)

Coffee Break

Wednesday, September 23 11:00 - 12:40 (Europe/Rome)

WM-L10: Space-based radar systems and missions

Room: Ch5
Chairs: Federica Bordoni (German Aerospace Center (DLR), Germany), Nertjana Ustalli (German Aerospace Center (DLR), Germany)
11:00 Investigation into the Weight Update Rate for Scan-On-Receive Beamforming
Felipe Queiroz de Almeida (German Aerospace Center (DLR) & Microwaves and Radar Institute, Germany); Marwan Younis (German Aerospace Center (DLR), Germany); Gerhard Krieger (DLR, Germany); Scott Hensley (Jet Propulsion Laboratory, USA); Alberto Moreira (German Aerospace Center (DLR), Germany)
Scan-On-Receive (SCORE) is a promising digital beamforming technique to improve the performance of SAR systems and enable high-resolution wide-swath concepts with multiple elevation beams. The technique is based on a beam-steering law which follows the angle-of-arrival of a pulse's echo traversing the ground. In practice, SCORE, also known as SweepSAR, is implemented on-board using digital processing, which implies discretized beam forming coefficients. This paper analyzes the implications of a beam switch to examine criteria for setting the necessary beam granularity and proposes a simple yet effective method to achieve pattern phase continuity.
11:20 Impact of Ambiguity Statistics on Information Retrieval for Conventional and Novel SAR Modes
Nertjana Ustalli and Michelangelo Villano (German Aerospace Center (DLR), Germany)
Synthetic aperture radar (SAR) is an all-weather imaging tool with several applications in the field of Earth observation. In recent years, novel SAR modes have been developed, where either the pulse repetition interval (PRI) or the transmitted waveform or phase is continuously varied, leading to smeared ambiguities. Ambiguities with the same total energy but different statistical distributions are likely to have different influences on the retrieval of information from SAR images. This paper analyzes the ambiguity statistics and addresses their impact on selected applications for conventional and novel SAR modes through simple models and simulations based on TerraSAR-X and TanDEM-X data. The reported results show that an understanding of the ambiguity statistics and of the impact of their smearing on product performance is fundamental to select the SAR operation mode within the design of future SAR systems.
11:40 Possible Sources of Imaging Performance Degradation in Advanced Spaceborne SAR Systems Based on Scan-on-Receive
Federica Bordoni (German Aerospace Center (DLR), Germany); Marc Rodriguez-Cassola and Gerhard Krieger (DLR, Germany)
The scan-on-receive capability is a fundamental feature of advanced spaceborne synthetic aperture radar (SAR) systems for high-resolution wide-swath imaging. The achievable SAR imaging performance depends on the possibility to properly collect the backscattered energy of interest, and to effectively suppress undesired returns. Nevertheless, the related signal model, reported in the literature, does not reflect the complexity of a real acquisition scenario. This paper identifies possible sources of image quality degradation, generally neglected, specific of this kind of systems. It provides a first analysis of their effect in dependence of instrument and acquisition geometry parameters, as support for tasks of system design and image quality assessment.
12:00 SPRATS: A Versatile Simulation and Processing RAdar ToolS for Planetary Missions
Oriane Gassot, Alain Hérique and Yves Rogez (Université Grenoble Alpes, CNRS, IPAG, France); Wlodek Kofman (Université Grenoble Alpes, CNRS, IPAG/Space Research Centre, PAS, France); Sonia Zine and Petit-prince Ludimbulu (Université Grenoble Alpes, CNRS, IPAG, France)
Our knowledge of the internal structure of asteroids is currently indirect and relies entirely on inferences from remote sensing observations of the surfaces. Radar observation of asteroids is the most mature technique available in order to characterize their structure, which is fundamental for understanding the small bodies' history and for planetary defense missions. However, as the small bodies' geometry is complex, simulation of the radar design and the signal is required in order to assess the instrument performance and thus the mission science return. SPRATS is a software package designed for mission and operation analysis for planetary bodies. It provides the tools to evaluate the radar concepts and mission scenarios, evaluate the processing performances, validate data processing techniques on the final data, and process real radar data when facing complex geometries different from the Earth observation geometry. The paper presents the architecture of the SPRATS toolbox along with relevant examples to illustrate its interest.
12:20 VISAR and VENSAR: Two Proposed Radar Investigations of Venus
Scott Hensley (Jet Propulsion Laboratory, USA); Bruce Campbell (Smithsonian National Air and Space Museum, USA); Dragana Perkovic (Jet Propulsion Laboratory, California Institute of Technology, USA); Kevin Wheeler (Jet Propulsion Laboratory, USA); Walter Kiefer (Lunar and Planetary Institute, USA); Richard Ghail (Royal Holloway, University of London, United Kingdom (Great Britain))
The last time scientists got a detailed view of the surface of Venus was with the revolutionary NASA Magellan Mission more than 30 years ago. Two new mission concepts under study by NASA and ESA are designed to return to Venus and obtain refined imagery and mapping data to answer some of the major questions concerning the evolution and current geologic activity of Venus. Both these missions would employ synthetic aperture radars to view the surface through the thick optically opaque atmosphere. Each radar design was optimized to meet the science objectives of their respective missions and bring back new data to the planetary science community to unravel why Venus, our twin planet, is so very different than the Earth. This paper discusses the complementary nature of these missions, the design and performance of these two radar instruments, and how each will further Venus and planetary science.

WM-L6: Compressed sensing and sparsity exploitation

Room: Ch1
Chairs: Moeness G. Amin (Villanova University, USA), Abdelhak M Zoubir (Darmstadt University of Technology, Germany)
11:00 Improved IF Estimation of Multi-Component FM Signals Through Iterative Adaptive Missing Data Recovery
Vaishali Amin and Yimin D. Zhang (Temple University, USA); Braham Himed (AFRL, USA)
In this paper, we address a challenging problem of accurate instantaneous frequency (IF) estimation of multi-component non-linear frequency modulated (FM) signals with distinct amplitude levels in the presence of missing data samples. In such scenarios, it is often difficult to resolve the weaker signal components. Besides, missing data-induced artifacts spread in the time-frequency (TF) domain, further complicating IF estimation. We propose a method that iteratively performs missing data recovery in the time-lag domain based on the least squares criterion in conjunction with signal-adaptive TF kernels. The proposed technique successfully resolves signal components with distinct amplitude levels, preserves a high resolution of the auto-terms and achieves robust TF distributions by mitigating the undesired effects of cross-terms and artifacts due to missing data samples. The effectiveness of the proposed method is verified through various simulation results.
11:20 Phase Transition in Frequency Agile Radar Using Compressed Sensing
Yuhan Li, Tianyao Huang, Xingyu Xu and Yimin Liu (Tsinghua University, China); Yonina C. Eldar (Weizmann Institute of Science, Israel)
Frequency Agile Radar (FAR) has improved anti-jamming performance over traditional pulse-Doppler radars under complex electromagnetic circumstances. To reconstruct the range-Doppler information in FAR, many compressed sensing (CS) methods including standard and block sparse recovery have been applied. In this paper, we study phase transitions of range-Doppler recovery in FAR using CS. In particular, we derive the closed-form phase transition curves associated with block sparse recovery and complex Gaussian matrices, based on prior results of standard sparse recovery under real Gaussian matrices. Empirical simulations validate the obtained curves, which lead to tight bounds on the radar and target parameters that guarantee exact recovery of targets' range-Doppler information.
11:40 Optimum Sparse Array Configuration for Mismatched DOA on Moving Platforms
Guodong Qin (Xidian University, China); Moeness G. Amin (Villanova University, USA)
Sparse arrays on a moving platform can efficiently expand the numbers of achievable degrees of freedom (DOFs) and consecutive lags, improving direction-of-arrival (DOA) estimation. This property has been recently used for the case of environment-independent structured sparse array configurations, such as those defined by nested and co-prime arrays. In this paper, we consider environment-dependent sparse arrays (EDSAs) design for mismatched DOA using Cramer-Rao bound (CRB) as the metric of optimality. The CRB is derived for multiple sources with angle biases. Two optimization metrics are constructed, where the enumeration is used to obtain the optimum arrays. Simulation results are provided to validate the effectiveness of the proposed EDSA design.
12:00 A Sparsity Based CFAR Algorithm for Dense Radar Targets
Jingxuan Chen (Xidian University, China); Shenghua Zhou (National Laboratory of Radar Signal Processing, Xidian University, China); Pramod Varshney (Syracuse University, USA); Jing Lu (Xidian University & National Laboratory of Radar Signal Processing, China); Jibin Zheng and Liu Hongwei (Xidian University, China); Hongtao Su (National Laboratory of Radar Signal Processing, China)
Conventional constant false alarm rate (CFAR) radar target detection algorithms, such as the cell-averaging (CA) CFAR and the order statistic (OS) CFAR, performs poorly with dense targets. In this paper, we develop a sparse signal processing-based CFAR algorithm for dense target detection. Under the \(\ell_1\)-norm constraint, we develop the relationship between the coefficient and the false alarm rate and verify the false alarm rate via numerical simulations. Numerical results indicate that the sparse signal processing-based method performs better in dense targets scenarios than CA-CFAR and OS-CFAR. The computation cost is evaluated both theoretically and numerically. We show that the computation time of MM-CFAR and CA-CFAR algorithms increases nearly linearly with the sample size.
12:20 Truncated Nuclear Norm Regularization Based Covariance Estimation for Airborne STAP
Ming Li (University of Electronic Science and Technology of China, China); Guohao Sun (Sichuan University, China); Zi-Shu He (University of Electronics Science and Technology of China, China)
The low-rank feature of clutter covariance matrix (CCM) can reduce the number of training samples required for space-time adaptive processing (STAP). Therefore, this paper proposes a novel covariance estimation method based on truncated nuclear norm regularization (TNNR) for airborne STAP in heterogeneous environment. Specifically, TNNR is introduced to ensure the low rank estimation of CCM and transform the NP-hard problem to be convex. Different from the conventional rank minimization algorithm, such as the nuclear norm relaxation approach, the proposed TNNR based approach only minimizes portion of the smallest singular values, which are unrelated to the focused subspace, deriving a more accurate approximation to the rank function. Moreover, we combine the low rank property with the block-Toeplitz structure to estimate the CCM. The block-Toeplitz structure constrain can ensure the model established in continuous domain without off-grid problem. The simulation results indicate that the proposed CCM estimation algorithm based on TNNR outperforms several other methods.

WM-SS10: Detection, tracking, and classification of small drones

Room: Ch4
Chairs: Carmine Clemente (University of Strathclyde, United Kingdom (Great Britain)), Francesco Fioranelli (TU Delft, The Netherlands)
11:00 A Feature-Based Approach for Loaded/Unloaded Drones Classification Exploiting micro-Doppler Signatures
Luca Pallotta (University of Roma Tre, Italy); Carmine Clemente (University of Strathclyde, United Kingdom (Great Britain)); Alessandro Raddi and Gaetano Giunta (University of Roma Tre, Italy)
This paper deals with the problem of loaded/unloaded drones classification. Precisely, exploiting the different micro-Doppler signatures exhibited by a drone with both any load and payloads of different weights, a novel signature extraction procedure is developed for automatic recognition purposes. The developed algorithms is based on a novel adaptation of the spectral kurtosis technique to the problem at hand, specifically the analysis of narrowband and wideband spectrograms of the radar echoes reflected by the drones. In addition, the principal component analysis is used to reduce the feature vector size. The experiments conducted on measured bistatic radar data prove the effectiveness of the proposed method in separating the quoted classes of objects.
11:20 On the Impact of Drone Airscrews Signature on Passive Radar Detection and Tracking Stages
Maria-Pilar Jarabo-Amores and David Mata-Moya (University of Alcalá, Spain); Pedro-Jose Gomez-del-Hoyo, Nerea del-Rey-Maestre and Javier Rosado-Sanz (University of Alcala, Spain)
In this paper, the impact of the drone airscrews signature in passive radars exploiting Digital Video Broadcasting signals is tackled. Beamforming techniques optimized to minimize the variance of the interference using the statistical characterization of real clutter data followed by a CFAR detector are used to detect drones and their associated airscrews signature. An independent bi-dimensional tracker based on the Kalman filter is proposed to tackle the drone and airscrews tracking process, and to extract micro-Doppler signature features. Real radar data acquired by a technological demonstrator developed at the University of Alcalá were used to evaluate the proposed solution. Results proved the suitability of the proposed detection and tracking schemes for airscrews signature feature extraction. Ansys HFSS electromagnetic simulator validated the results.
11:40 Exploitation of Long Coherent Integration Times to Improve Drone Detection in DVB-S Based Passive Radar
Tatiana Martelli (Sapienza University of Rome, Italy); Octavio Cabrera (Sapienza University, Italy); Fabiola Colone (Sapienza University of Rome, Italy); Pierfrancesco Lombardo (University of Rome La Sapienza, Italy)
In this paper, we consider the exploitation of long coherent integration times for improving the detection of small drones in a DVB-S based Passive Radar (PR). In fact, since these low-flying objects are characterized by very small RCS values, their detection is challenging especially when using DVB-S signals as source of opportunity, typically received with low power levels. In this work, in order to benefit from the use of extended coherent integration times, the target range and Doppler migration effects have been properly compensated. Then, the effectiveness of the adopted processing scheme has been tested against real data collected by a DVB-S based PR receiver developed at Sapienza University of Rome. The reported results show that the coherent integration times can be increased up to a few seconds thus improving the drone detection capability of the considered system.
12:00 Use of Symmetrical Peak Extraction in Drone Micro-Doppler Classification for Staring Radar
Cameron Bennett (University oif Glasgom, United Kingdom (Great Britain)); Mohammed Jahangir (University of Birmingham, United Kingdom (Great Britain)); Francesco Fioranelli (TU Delft, The Netherlands); Bashar I. Ahmad (University of Cambridge, United Kingdom (Great Britain)); Julien Le Kernec (University of Glasgow, United Kingdom (Great Britain))
The commercialization of drones has granted the public with unprecedented access to unmanned aviation. As such, the detection, tracking, and classification of drones in radars have become an area in high demand to mitigate accidental or voluntary misuse of these platforms. This paper focuses on the classification of drone targets in a safety context where the concept of Explainable AI is of particular interest. Here, we propose a simple, yet effective, means to extract a salient symmetry feature from the micro-Doppler signatures of drone targets, arising from onboard rotary components. Most importantly, this approach maintains the explainable nature of the employed recognition algorithm as the symmetry feature is directly related to the kinematics of the drones as the targets of interest. A large dataset collected from multiple locations with over 280 minutes of rotary and fixed wing drone flights has been collected and used to demonstrate the generalization capability of this approach.
12:20 Radar Signatures of Drones Equipped with Liquid Spray Payloads
Samiur Rahman (University of St Andrews, United Kingdom (Great Britain)); Duncan Robertson (The University of St Andrews, United Kingdom (Great Britain)); Mark Govoni (US Army Research Laboratory, USA)
The widespread availability of cheap and robust commercial drones has increased the likelihood of these being used for malicious purposes. In some cases they may be equipped with threat payloads. This study reports on the distinctive radar signatures of drones spraying liquid, analogous to a drone delivering chemical weapons, for example. A commercially available crop spraying drone has been used as the basis for liquid droplet radar backscatter modelling and for experimental data acquisition. The spray nozzle droplet parameters were used to model the radar cross section (RCS) and the signal-to-noise ratio (SNR) of the liquid droplets at X-, K- and W-bands, using the Rayleigh approximation. Additionally, experimental data have been obtained simultaneously with 24 GHz and 94 GHz radars. The processed results show that they are in very good agreement with the model. It is clearly demonstrated that at W-band (94 GHz), the liquid spray produces strong micro-Doppler signatures observed from the range-Doppler plots whereas no such detection was possible at K-band (24 GHz). The experimental results validate the hypothesis that millimeter-wave radar offers superior sensitivity than lower frequency bands to reflections from liquid spray droplets of <<0.5 mm size. Hence, a millimeter-wave radar system can potentially be used for classifying a drone with a liquid spray payload.

WM-SS8: Coexistence of radar and communication systems

Room: Ch2
Chair: Stefano Fortunati (CentraleSupélec, France)
11:00 A Power Control Problem for a Dual Communication-Radar System Facing a Jamming Threat
Andrey Garnaev (WINLAB, Rutgers University, USA); Athina Petropulu (Rutgers, The State University of New Jersey, USA); Wade Trappe (WINLAB, Rutgers University, USA); H. Vincent Poor (Princeton University, USA)
A dual purpose communication and radar system is considered, operating in the presence of a jammer. The system transmits communication signals and uses their reflections off targets to support target tracking. A game theoretic framework is used to design the system, i.e., determine power allocations, so that the performance of both communication and radar components is optimized. Two metrics are considered and compared for system design: (a) a weighted combination of signal-to-interference-plus-noise ratio (SINR) of the communication and radar components, reflecting the performance of both components; and (b) a weighted combination of the inverse SINRs of the corresponding components, reflecting the latency in the performance of the components. The comparison suggests that the latency metric enables a design that requires less a priori information on fading gains, and further, as the decision rule can be derived in closed form, the design is easier implementable in practice. The interaction between the system and the jammer is modeled by non-zero sum games. A Bayesian extension of the games allows us to investigate the impact that incomplete information about the underlying network parameters has upon the strategies that should be employed.
11:20 Automotive Radar Interference Characterization and Reduction by Partial Coordination
Khurram Usman (The University of Texas at Austin, USA); Robert Heath (North Carolina State University & The University of Texas at Austin, USA); Kapil Gulati and Junyi Li (Qualcomm, USA)
In this paper, we characterize the system level effect of radar interference in terms of the probability of missed detection and the probability of false alarm for frequency modulated continuous wave (FMCW) radar sensors in a highway type automotive setting. Our results show that the probability of missed detection is significantly affected by the interference, particularly at larger distances. We also propose a central agent that uses the existing communication infrastructure to jointly assign different FMCW parameters, such as slope directions and carrier frequency offsets, to the FMCW radar units operating in its vicinity. Results from our system level simulation show that the proposed resource allocation strategy outperforms the widely reported solution to randomize the radar parameters for interference reduction.
11:40 Joint Beamforming Design for Extended Target Estimation and Multiuser Communication
Fan Liu and Christos Masouros (University College London, United Kingdom (Great Britain))
In this paper, we propose a novel multi-input multi-output (MIMO) beamforming design towards joint radar sensing and multiuser communication. To characterize the performance of the radar functionality, we derive a closed form of the mean squared error (MSE) for target estimation given a probing signal. Furthermore, we propose to minimize the estimation MSE while guaranteeing the signal-to-interference-plus-noise ratio (SINR) of each communication user. While the formulated optimization problem is non-convex, it can be solved via the semidefinite relaxation algorithm. Finally, numerical results are provided to validate the effectiveness of the proposed approach.
12:00 Joint Radar Target Detection and Parameter Estimation with MIMO OTFS
Lorenzo Gaudio (University of Parma, Italy); Mari Kobayashi (CentraleSupelec, France); Giuseppe Caire (Technische Universität Berlin, Germany); Giulio Colavolpe (University of Parma, Italy)
Motivated by future automotive applications, we study the joint target detection and parameter estimation problem using orthogonal time frequency space (OTFS), a digital modulation format robust to time-frequency selective channels. Assuming the transmitter is equipped with a mono-static MIMO radar, we propose an efficient maximum likelihood based approach to detect targets and estimate the corresponding delay, Doppler, and angle-of-arrival parameters. In order to reduce the computational complexity associated to the high-dimensional search, our scheme proceeds in two steps, i.e., target detection and coarse parameter estimation followed by refined parameter estimation. Interestingly, our numerical results demonstrate that the proposed scheme is able to identify multiple targets if they are separated in at least one domain out of three (delay, Doppler, and angle), while achieving the Cram\'er-Rao lower bound for the parameter estimation.
12:20 On Wilks' Theorem for Generalized Likelihood Ratio Test Performance of Cooperative Radar-Communications
Touseef Ali and Akshay Bondre (Arizona State University, USA); Christ D. Richmond (Arizona State University & Ira A. Fulton Schools of Engineering, USA)
The analysis of the generalized likelihood ratio test (GLRT) radar receiver when the radar is coexisting with a cooperative in-band communication (comm.) system is challenged by the discrete nature of the comm. symbols. While the asymptotic performance of the GLRT can be well-approximated by using the Fisher information matrix (FIM) for the Cram´er-Rao bound (CRB) via Wilks' theorem, this only applies when the underlying data distribution for each hypothesis is continuous in the parameters. Thus, to accommodate the discrete nature of the comm. symbols an adaptation to Wilks' theorem is proposed in this work that uses the constrained CRB (CCRB) in place of the usual unconstrained CRB. This development is formulated with the complexified CCRB first, and then mapped to an equivalent real FIM for use in Wilks' result.

WM-SS9: Interference in automotive radars

Room: Ch3
Chairs: Kumar Vijay Mishra (United States Army Research Laboratory, USA), Bhavani Shankar Mysore R (Interdisciplinary Centre for Security, Reliability and Trust & University of Luxembourg, Luxembourg)
11:00 Comparison of Automotive FMCW and OFDM Radar Under Interference
Gisela Carvajal (Qamcom Research and Technology, Sweden); Musa Furkan Keskin and Canan Aydogdu (Chalmers University of Technology, Sweden); Olof Eriksson and Hans Herbertsson (Veoneer, Sweden); Hans Hellsten (Saab Microwave Systems, Sweden); Emil Nilsson (Halmstad University, Sweden); Mats Rydström (Veoneer, Sweden); Karl Vänas (Volvo Cars, Sweden); Henk Wymeersch (Chalmers University of Technology, Sweden)
Automotive radars are subject to interference in spectrally congested environments. To mitigate this interference, various waveforms have been proposed. We compare two waveforms (FMCW and OFDM) in terms of their radar performance and robustness to interference, under similar parameter settings. Our results indicate that under proper windowing both waveforms can achieve similar performance, but OFDM is more sensitive to interference.
11:20 Stochastic-Geometry-Based Interference Modeling in Automotive Radars Using Matérn Hard-Core Process
Kumar Vijay Mishra (United States Army Research Laboratory, USA); Bhavani Shankar Mysore R (Interdisciplinary Centre for Security, Reliability and Trust & University of Luxembourg, Luxembourg); Björn Ottersten (University of Luxembourg, Luxembourg)
As the use of radars in autonomous driving systems becomes more prevalent, these systems are increasingly susceptible to mutual interference. In this paper, we employ stochastic geometry to model the automotive radar interference in realistic traffic scenarios and then derive trade-offs between the radar design parameters and detection probability. Prior works model the locations of radars in the lane as a homogeneous Poisson point process (PPP). However, the PPP models assume all nodes to be independent, do not account for the lengths of vehicles, and ignore spatial mutual exclusion. In order to provide a more realistic interference effect, we adopt the Matérn hard-core process (MHCP) instead of PPP, in which two vehicles are not closer than an exclusion radius from one another. We show that the MHCP model leads to more practical design trade-offs for adapting the radar parameters than the conventional PPP model.
11:40 Sweep-Based Spectrum Sensing Method for Interference-Aware Cognitive Automotive Radar
Gor Hakobyan and Martin Fink (Robert Bosch GmbH, Germany); Ayinhu Soyolyn, Nour Mansour and Dirk Dahlhaus (University of Kassel, Germany)
The ongoing automation of driving functions in cars leads to a massive growth in the number of automotive radar sensors, and thus to more radar interference. An approach for actively mitigating interference between automotive radars is the interference-aware cognitive radar (IACR). One major challenge for IACR is, however, sensing of a large spectral band (e.g. 77-81 GHz) potentially available for radar operation. In this paper, we present a cost-efficient spectrum sensing method based on linear sweeping over a wide spectral band. The signal resulting from the sweep is lowpass filtered prior to sampling, which allows a significant bandwidth reduction down to few tens of MHz. In the digital domain, the captured signal is pulse-compressed with a bank of matched filters. The output image of the time-frequency space provides a basis for identification of interference-free regions and a subsequent adaptation for the next measurement cycle. The performance of the proposed approach is studied in simulation and demonstrated with a prototype. The results indicate the feasibility of such spectrum sensing module for automotive radar both in terms of performance and cost.
12:00 Automotive Radar and MmWave MIMO V2X Communications: Interference or Fruitful Coexistence
Andrew Graff (The University of Texas at Austin, USA); Anum Ali (University of Texas at Austin, USA); Nuria González-Prelcic (North Carolina State University, USA); Amitava Ghosh (Nokia Bell Labs, USA)
Radar-aided millimeter wave (mmWave) vehicular communication can efficiently reduce the training overhead associated with configuring/reconfiguring the large antenna arrays used to operate at these frequencies. Previous work neglects, however, the effect of the interference of the radar signal into the communication receiver. In this paper, we consider a cellular-supported vehicle-to-infrastructure millimeter wave system with a base station-mounted radar used to speed up the beam training process. We develop a measurement setup to evaluate the radar interference levels that appear at the communication receiver in the vehicle. We also implement a simulation of this vehicular system that can be used to obtain performance results that account for the effect of the radar interference. The spectral efficiency results show that the radar-aided vehicular communication system at mmWave does not suffer from a significant impact of the radar-to-communication interference, even when the car is close to the base station-mounted radar.
12:20 Combined Object Detection and Tracking on High Resolution Radar Imagery for Autonomous Driving Using Deep Neural Networks and Particle Filters
Ana Stroescu and Liam Y. Daniel (University of Birmingham, United Kingdom (Great Britain)); Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain))
This paper presents a novel approach for target detection in radar imagery, which combines an object detector and a multi target particle filter tracker. Object detection is implemented using deep neural networks, as opposed to the traditional radar object detection methods. This technique is applied to a dataset collected with a 79 GHz FMCW radar mounted on a vehicle. In this approach, object detection and tracking of roadside objects are performed in an alternating fashion to reduce the computational load required by the real time processing. The results and the thorough analysis of the parameters showed that this approach is feasible and can be successfully utilised in radar imagery for autonomous driving.

Wednesday, September 23 12:40 - 2:00 (Europe/Rome)

Lunch Break

Wednesday, September 23 2:00 - 3:00 (Europe/Rome)

Plenary Talk 2 - Markus Gardill (University of Würzburg. Germany): State-of-the-Art Automotive Radar System Architectures - and What Else We Can Do with Them.

Chair: Antonio De Maio (University of Naples "Federico II", Italy)

Wednesday, September 23 3:00 - 4:40 (Europe/Rome)

WA-L3: Automotive radars

Room: Ch3
Chairs: Carlo Bongioanni (Sapienza University of Rome, Italy), Christian Waldschmidt (University of Ulm, Germany)
3:00 Monostatic RCS Measurements of Representative Road Traffic Objects in the 76 … 81 GHz Frequency Band
Sreehari Buddappagari Jayapal Gowdu, Andreas Schwind and Ralf Stephan (Technische Universität Ilmenau, Germany); Matthias Hein (Ilmenau University of Technology, Germany)
Quantitative knowledge of the radar cross-section of road traffic objects is a critical input for the design and performance evaluation of automotive radar systems in terms of target detection, discrimination, and tracking. Moreover, it has to be emulated accurately when applied in an over-the-air/vehicle-in-the-loop (OTA/ViL) based verification and validation method. Therefore, it is essential to measure RCS patterns of traffic objects under conditions with separable influence of ground reflections and minimized background artefacts. Given that in publicly available literature there are few reports of high-quality RCS data of representative targets like passenger car, bicycle and powered two-wheeler (PTW), measured over the entire frequency band between 76 GHz and 81 GHz, we have performed wideband RCS measurements for both horizontal and vertical co-polarizations. The measurements were performed in a semi-anechoic chamber and analysed with advanced post-processing for improving the accuracy of the measured data. Additional post-processing in time domain was performed to obtain two-dimensional contour plots of scattering locations. Results are presented in three data domains: frequency, statistical distribution, and time-domain 2D contour images, to deliver data for improved radar-based target classification. The impact of polarization is also discussed.
3:20 Performance Evaluation of Wide Aperture Radar for Automotive Applications
Oded Bialer (University of Tel-Aviv, Israel); Sammy Kolpinizki (Tel Aviv University & General Motors ATCI, Israel); Amnon Jonas (General Motors, Israel)
The angular resolution of the state-of-the-art automotive radars is about 1 degree, due to the relatively short aperture of about 10cm. In this paper we present novel results of a 1m aperture automotive prototype radar that achieves 0.1 degree resolution. Since the 1m aperture is about the maximal horizontal aperture of a vehicle, then 0.1 degree angular resolution is the maximal practical resolution with 78GHz technology. Hence the demonstrated performance can be considered as the upper performance limit of the 78GHz frequency. We analyze the wide aperture radar performance advantage with respect to state-of-the-art short aperture radar and high resolution LIDAR sensor, in various challenging automotive scenarios. The wide aperture radar attains a relatively large number of detection points from the object surface, which are much more accurate than the conventional short aperture radar, and therefore enables to accurately estimate the object position, size and boundaries, which is essential for autonomous driving. It attains comparable performance to LIDAR at short to medium range, and outperforms LIDAR at long range. The pioneer performance evaluation presented in this paper shows the high potential performance of wide aperture automotive radars and motivates their application in next generation autonomous driving systems.
3:40 Compressive Sensing for Automotive 300GHz 3D Imaging Radar
Dominic Michael Phippen (University of Birmingham & Cranfield University, United Kingdom (Great Britain)); Liam Y. Daniel, Edward Hoare and Mikhail Cherniakov (University of Birmingham, United Kingdom (Great Britain)); Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain))
Novel 3D radar imaging techniques exploiting extremely wide band signals in the low-THz (0.1-1 THz) radar band is proposed and experimentally verified in this paper. This paper provides comparative analysis of the abilities of three imaging techniques with varying levels of sophistication: bilateration, backprojection, and elastic net. These techniques are applied to a 300GHz radar and results are discussed.
4:00 Radar Tracking with Orthogonal Velocity Measurements for Autonomous Ground Vehicles
Eric Klinefelter, Jeffrey Nanzer and Hayder Radha (Michigan State University, USA)
In this work we examine the effect of having independent radial and angular velocity radar measurements on tracking performance. With this, we provide design considerations for systems using interferometric angular velocity estimation. We found that on average as the number of clutter detections increases, the reduction in error by adding an additional angular velocity measurement also increases, implying a greater need for orthogonal velocity measurements in high-clutter environments. In particular, in simulation at a clutter rate of 50 detections per measurement, we obtained an error reduction of 14% by adding a radial velocity measurement and a further reduction of 3% from an additional angular velocity measurement. We also tested this method on a publicly available autonomous driving dataset by synthesizing angular velocity measurements and found a similar reduction of 13% and 6% by adding Doppler and angular velocity measurements, respectively.
4:20 Extended Object Tracking with Automotive Radar Using Learned Structural Measurement Model
Yuxuan Xia (Chalmers University of Technology, Sweden); Pu Wang (Mitsubishi Electric Research Laboratories (MERL), USA); Karl OE Berntorp (Mitsubishi Electric Research Labs, USA); Petros T. Boufounos (Mitsubishi Electric Research Laboratories & Rice University, USA); Philip Orlik (Mitsubishi Electric Research Laboratories, USA); Lennart Svensson (Chalmers University, Sweden); Karl Granström (Chalmers University of Technology, Sweden)
This paper presents a data-driven measurement model for extended object tracking (EOT) with automotive radar. Specifically, the spatial distribution of automotive radar measurements is modeled as a hierarchical truncated Gaussian with structural geometry parameters (e.g., truncation bounds, their orientation, and a scaling factor) learned from the training data. The contribution is twofold. First, the learned measurement model can provide an adequate resemblance to the spatial distribution of real-world automotive radar measurements. Second, large-scale offline training datasets can be leveraged to learn the geometry-related parameters and offload the computationally demanding model parameter estimation from the state update step. The learned structural measurement model is further incorporated into the random matrix-based EOT approach with a new state update step. The effectiveness of the proposed approach is verified on the nuScenes dataset.

WA-L4: Radar for structural monitoring

Room: Ch4
Chairs: Guido Manacorda (IDS GeoRadar, Italy), Lapo Miccinesi (University of Florence, Italy)
3:00 Extraction of Small Objects from Ground-Based Multi-Static SAR Images Using CFAR Algorithm with Generalized Gamma Distribution
Andreas Heinzel, Eric Schreiber, Markus Peichl and Stephan Dill (German Aerospace Center (DLR), Germany)
Ground-based multi-static synthetic aperture radar is able to produce high-resolution images with sufficient clutter suppression. It is therefore capable of detecting even small and weakly reflecting objects. The automatic extraction of these targets is of great interest for many applications like efficient landmine detection, for instance. Furthermore the automatic extraction of targets is the first useful stage for automatic target recognition. A Constant False Alarm Rate (CFAR) algorithm is a suitable approach for this task by analyzing the distribution of the surrounding clutter and calculating a reasonable threshold for each pixel. The Generalized Gamma distribution forms a large class of distributions and is therefore suitable to model different kind of clutter. In this paper the selected and implemented algorithm is validated using near-range high-resolution data from different kind of objects.
3:20 Copula-Based Robust Landmine Detection in Multi-View Forward-Looking GPR Imagery
Afief Dias Pambudi (Signal Processing Group, Technische Universität Darmstadt, Germany); Fauzia Ahmad (Temple University, USA); Abdelhak M Zoubir (Darmstadt University of Technology, Germany)
We propose a scheme for detecting landmines using forward-looking ground-penetrating radar. The detector is applied to tomographic radar images obtained from multiple viewpoints of the investigation area and is based on a robust version of the likelihood-ratio test. The statistical dependence between multi-view images is incorporated via a copula-based function. The test is designed to maximize the worst-case performance over all feasible target and clutter distribution pairs, thereby eliminating the need for a strong assumption about the clutter distribution. Using numerical radar data of shallow buried targets, we demonstrate the superior performance of the proposed detector over existing approaches.
3:40 A Spatial Inference Approach for Landmine Detection Using Forward-Looking GPR
Afief Dias Pambudi (Signal Processing Group, Technische Universität Darmstadt, Germany); Martin Gölz (Darmstadt University of Technology, Germany); Fauzia Ahmad (Temple University, USA); Abdelhak M Zoubir (Darmstadt University of Technology, Germany)
We propose to detect landmines and unexploded ordnance in forward-looking ground-penetrating radar imagery by applying a spatial multiple hypothesis testing method. Homogeneous regions in an investigation area are identified based on spatial proximity and similarity of the observed decision statistics in order to discriminate targets against clutter. The proposed method is designed to control the proportion of false alarms among all pixels declared to be associated with a target. The detection performance of the proposed method is evaluated using numerical data of shallow-buried targets and compared to existing approaches.
4:00 Comparison Between Sparse Array and Compressive Sensing for Designing a 4X4 MIMO Radar
Massimiliano Pieraccini, Lapo Miccinesi and Enrico Boni (University of Florence, Italy)
Multiple Input Multiple Output (MIMO) radar allows to reduce the number of antennas maintaining performances comparable to a phased array radar. Sparse array (SA) and Compressive Sensing (CS) are two different approaches that promise to reduce further the number of antennas. The aim of this paper is to compare the two different approaches for designing a 4x4 MIMO. The simulated results showed that side lobe level (SLL) for CS is lower than that of SA by at least 15 dB. Nevertheless, the authors note that SA pattern could be more properly optimized with a simulated annealing that could lower a little the side lobes. Moreover, all the simulations have been conducted with a single point target, but a comprehensive evaluation with more complex scenarios, such as multiple targets and/or spread targets, should be conducted as well.
4:20 RockSpot: An Interferometric Doppler Radar for Rockfall/Avalanche Detection and Tracking
Federico Viviani, Alberto Michelini and Lorenzo Mayer (IDS GeoRadar Srl, Italy)
In the last decade ground-based SAR (GB-SAR) systems has been demonstrated to be one of the most efficient tools for slope monitoring in mountainous regions, costal riffs, quarries, and open pit mines. These systems can measure small displacements at a distance for terrain back scattering targets but are not able to deal with fast moving targets, such as rockfall events or rock/snow avalanches. In this paper, an innovative interferometric doppler radar system able to locate and track rockfalls and avalanches in real-time is presented.

WA-L5: Space-based/Airborne radar remote sensing applications

Room: Ch5
Chair: Gianfranco Fornaro (CNR-IREA, Italy)
3:00 A Fast Forward-looking Super-resolution Imaging Method for Scanning Radar Based on Low-rank Approximation with Least Squares
Xingyu Tuo, Yin Zhang and Yulin Huang (University of Electronic Science and Technology of China, China)
In forward-looking imaging of scanning radar, high range resolution can be realized by matured approaches, while poor angular resolution restricts application of the scanning radar. Several super-resolution methods have been proposed to improve the angular resolution, but they ignore the redundancy characteristics of the antenna measurement matrix. In this paper, we propose a fast forward-looking super-resolution imaging method for scanning radar based on low-rank approximation with least squares. To reduce the redundancy, the low-rank approximation method is used to extract the main information of the antenna measurement matrix and echo matrix, the least squares method is employed to solve the objective function. Simulations and experiment prove the proposed method can promote computational efficiency without losing super-resolution performance.
3:20 Airborne Passive Radar Detection for the APART-GAS Trial
Vichet Duk and Philipp M. Wojaczek (Fraunhofer FHR, Germany); Luke Rosenberg (Defence, Science and Technology Group & University of Adelaide, Australia); Diego Cristallini and Daniel W O'Hagan (Fraunhofer FHR, Germany)
In recent years, passive radar has been investigated for airborne applications in imaging and detection of ground targets. There are challenges with detection of land targets from an airborne platform due to platform motion which causes the clutter Doppler to spread and the presence of discrete scatterers which can be strong and confused with targets of interest. In this paper, we compare two different detection schemes to assess their performance on a real airborne passive radar data from a land scene North of Poland. These include a standard coherent scheme and a sparse signal separation algorithm known as morphological component analysis. The performance is measured using a Monte Carlo simulation where synthetic targets are injected into the real data set.
3:40 A Novel Geo-Statistical Approach for Transport Infrastructure Network Monitoring by Persistent Scatterer Interferometry (PSI)
Valerio Gagliardi, Luca Bianchini Ciampoli and Fabrizio D'Amico (Roma Tre University, Italy); Fabio Tosti (University of West London, United Kingdom (Great Britain)); Amir M. Alani (University of West London (UWL), United Kingdom (Great Britain)); Andrea Benedetto (Roma Tre University, Italy)
Persistent Scatterer Interferometry (PSI) is an Interferometric Synthetic Aperture Radar (InSAR) remote sensing technique based on a multi-temporal interferogram analysis of SAR images. The aim of the technique is to extract long-term high phase stability benchmarks of coherent point targets, namely Persistent Scatterers (PSs). In the last few years, several approaches have been developed to obtain PSI point targets, proving their viability for applications to transport infrastructure monitoring and surveillance. However, SAR satellites can only detect displacements in the Line-of-Sight (LoS), with reference to the specific orbit-related incident angle. This work proposes a novel geo-statistical approach to ease the post-processing of large datasets of PSs resulting from the application of the PSI algorithms over an area of interest. The approach aims at correcting the component of the displacement collected from the acquisition geometry of the sensor. Keywords-Persistent Scatterers Interferometry (PSI), Transport Infrastructure Maintenance, acquisition geometry, geo-statistical approach, Line of Sight (LoS)
4:00 Multi-Channel Calibration for Airborne Post-Doppler Space-Time Adaptive Processing
Andre Barros Cardoso da Silva (German Aerospace Center (DLR), Germany); Stefan V. Baumgartner (German Aerospace Center (DLR), Germany)
This paper presents a fast and efficient multi-channel calibration algorithm for along-track systems, which in particular is evaluated for the post-Doppler space-time adaptive processing (PD STAP) technique. The calibration algorithm corrects the phase and magnitude offsets among the receiving channels, estimates and compensates the Doppler centroid variation caused by atmospheric turbulences by using the attitude angles of the antenna array. Important parameters and offsets are estimated directly from the radar range-compressed data. The proposed algorithm is compared with the state-of-the-art Digital Channel Balancing technique based on real multi-channel X-band data acquired by the DLR's airborne system F-SAR. The experimental results are shown and discussed in the frame of traffic monitoring applications.
4:20 Drone-Borne P-band Single-Pass InSAR
Laila Moreira (Radaz, Brazil); Dieter Lübeck (Radaz Ltda, Brazil); Christian Wimmer (Wimmer Consulting, Germany); Felicio Harley Garcia de Castro (UNICAMP, Brazil); Juliana A Góes (University of Campinas - UNICAMP, Brazil); Valquiria Lima Bessa de Castro, Marlon Alcântara and Gian Oré (University of Campinas, Brazil); Luciano P. Oliveira (State University of Campinas, Brazil); Leonardo Sant´Anna Bins (INPE, Brazil); Barbara Teruel and Lucas H Gabrielli (University of Campinas, Brazil); Hugo Enrique Hernandez-Figueroa (Unicamp, Brazil)
This paper presents a drone-borne high-accuracy single-pass Interferometric Synthetic Aperture Radar System in the P-band (P-InSAR) for forest inventory, where ground and canopy heights are accurately determined. Full penetration is proven for a eucalyptus forest with a tree spacing of 2.5 m x 3.0 meters, and the measured digital terrain accuracy is compared with well-known statistical models. Combining a C-band single-pass InSAR with P-InSAR forest height is estimated with 5% accuracy for forest inventory. Both surface and ground digital models are presented and compared with ground truth measurements.

WA-SS11: Emerging technologies for radar applications: compressive sensing meets machine learning

Room: Ch1
Chairs: Laura Anitori (TNO, The Netherlands), Michael Kohler (Fraunhofer FHR, Germany)
3:00 A Fast-Learning Sparse Antenna Array
Satish Mulleti (Weizmann Institute of Science, Israel); Chiranjib Saha and Harpreet S Dhillon (Virginia Tech, USA); Yonina C. Eldar (Weizmann Institute of Science, Israel)
Selecting a sparse subset of antennas to obtain high-resolution direction-of-arrival estimates while circumventing the complexity associated with using a large array is critical in many radar applications. Since this subset selection problem is combinatorial, deep learning has been recently proposed as a possible solution for efficiently solving it. However, the bottleneck in this approach is training data generation, which requires an exhaustive search over all possible subarrays. In this paper, we propose an efficient method for generating training data using ideas from submodular optimization. In particular, we use the log-determinant of the Cramer-Rao lower bound as our cost function due to its submodular structure. It is then minimized through a greedy optimization approach to determine the best subarray. We provide numerical simulations to validate the performance of the proposed array selection strategy. Our simulations show that the proposed approach is ten times faster in training than an exhaustive search method while providing comparable performance.
3:20 Sparse Signal Models for Data Augmentation in Deep Learning ATR
Tushar Agarwal, Nithin Sugavanam and Emre Ertin (The Ohio State University, USA)
Automatic Target Recognition (ATR) algorithms classify a given Synthetic Aperture Radar (SAR) image into one of the known target classes using a set of training images. Recently, learning methods have shown to achieve state-of-the-art classification accuracy if abundant training data is available sampled uniformly over the classes and their poses. In this paper, we consider the problem of improving the generalization performance of learning methods in SAR-ATR when training data is limited. We propose a data augmentation approach using sparse signal models that capitalizes on commonly observed phenomenology of wide-angle synthetic aperture radar (SAR) imagery. Specifically, we exploit the sparsity of the scattering centers in the spatial domain as well as the limited persistence of the scattering coefficients in the azimuthal domain to solve the ill-posed problem of over-parametrized model fitting. Using this fitted model, we synthesize new images at poses not available in training set to augment the training data used by CNN. We validate the performance of the proposed model based data augmentation strategy on subsampled versions of the MSTAR dataset. The experimental results show that for the training data starved region, the proposed method provides a significant gain in the generalization performance of the resulting ATR algorithm.
3:40 Online Dictionary Learning Techniques for Sea Clutter Suppression
Fabio Giovanneschi (Fraunhofer FHR, Germany); Luke Rosenberg (Defence, Science and Technology Group & University of Adelaide, Australia); Diego Cristallini (Fraunhofer FHR, Germany)
Maritime airborne surveillance radars operating at high grazing angles experience significant sea clutter returns with a time and range-varying Doppler spectra, thus making the detection of small targets extremely difficult. Many target detection schemes assume models for the amplitude distribution that require estimation of specific parameters. These may not be accurate in all situations and hence reduce the detection performance. In this paper, we propose a new sea clutter suppression approach based on a dictionary learning strategy. The success of this data-driven approach is that it does not rely on statistical modeling of the clutter, but rather on the quality of the training data used to form the dictionary. Our analysis looks at the learning performance between batch and online strategies and then uses the obtained dictionaries to assess the clutter suppression capabilities in terms of improvement in the signal to interference ratio and the detection performance using a Monte Carlo analysis. To demonstrate these algorithms, data from the Australian Ingara L-band radar system has been used.
4:00 Combined CS and DL Techniques for DOA with a Rotman Lens
Matthias Weiß, Michael Kohler, Alexander Saam and Josef Worms (Fraunhofer FHR, Germany)
Rotman lenses are useful devices commonly utilized within multi-beam antenna array networks. They are generally used in radar surveillance systems to detect targets in multiple directions simultaneously without physically moving the antenna front-end. Nowadays, the communications sector (5G) also has great interest in this technology. Due to the use of a free-space true time delay network, for instance attached to Uniform Linear Array (ULA) of broadband Vivaldi antenna elements, this type of microwave lens support low-phase error, wide frequency band operation, and wide-angle scanning combined with simultaneous spatial beams (beamspace). In particular the multi-beam feature makes the lens very attractive for Direction of Arrival (DoA) estimation. This paper combines the aforementioned advantage with a dedicated Neural Network (NN) for an efficient wideband Direction of Arrival (DoA) and frequency estimation technique based on a single snapshot from such a multi-beam antenna configuration. The proposed approach uses machine learning techniques to establish the NN with a training set obtained from measurements in an anechoic chamber enriched superimposing different noise levels. This results in a lower computational load during the training phase and finally a very fast estimation of direction and frequency of the impinging signal.
4:20 Simultaneous Resolution of Range-Doppler Ambiguities Using Agile Pulse Intervals with Sparse Signal Processing
Wim Lambertus van Rossum and Laura Anitori (TNO, The Netherlands)
In this article a novel waveform based on agile (or irregular) pulse intervals is presented as a potential approach to simultaneously resolve both range and Doppler ambiguities using a single burst in radar measurements containing very far away and/or very fast targets. The paper shows that, the combination of irregularly spaced transmitted pulses with Sparse Signal Processing (SSP) on receive can be used to resolve radar ambiguous measurements within a single burst, in contrast to the more conventional PRF Staggering approach which uses multiple bursts, each with a different pulse repetition frequency. In addition to reducing the total time needed to measure the target parameters unambiguously, the proposed waveform offers also some robustness against certain types of Electronic Counter Measures. The claims of the paper are validated by means of simulations as well as with a set of experimental measurements that were carried out at the University of Warsaw, which preliminarily confirm our findings

WA-SS12: Dual-function radar/communication systems

Room: Ch2
Chairs: Shannon D Blunt (University of Kansas, USA), Aboulnasr Hassanien (Wright State University, USA)
3:00 On Spectrum Sharing for Pulse-Doppler Radar and OFDM Communications
Justin Metcalf (University of Oklahoma, Italy); Shane Flandermeyer (University of Oklahoma, USA)
Radar receivers must be extremely sensitive to detect faint target signatures. To this end pulse-Doppler processing of pulse compression waveforms integrates faint target echoes across multiple pulses by matching to the Doppler shift of target returns. Therefore, the combination of receiver sensitivity and exploitation of the target return frequency structure implies that radar signal processing will be sensitive to structured interference, such as that from communications systems. To explore this sensitivity, we examine the impact of a popular communications waveform, orthogonal frequency division multiplexing (OFDM), on the output of pulse-Doppler signal processing. Understanding the impact of cross-function interference will improve requirements generation for both dual-function and shared spectrum systems.
3:20 Assessment of Constant Envelope OFDM as a Class of Random FM Radar Waveforms
Erik Biehl (University of Kansas, USA); Charles Mohr (Air Force Research Laboratory); Brandon Ravenscroft and Shannon D Blunt (University of Kansas, USA)
Random FM waveforms designed to possess a desirable spectral shape (on average) have been experimentally demonstrated for an increasing number of applications. However, the benefits of these waveforms can be offset by the computational cost of performing real-time spectral-shaping optimization for each random initial waveform. Here the constant-envelope orthogonal frequency division multiplexing (CE-OFDM) framework from communications is examined as a means to generate random FM waveforms. This scheme is attractive because it inherently provides useful spectrum shaping without the need for optimization, thereby realizing an effective form of random FM that can be produced in real-time on systems with modest computational resources. Performance is assessed in simulation and using free-space experimental measurements.
3:40 A Study on the Performance of Symbol Dictionary Selection for the Frequency Hopped DFRC Scheme
William Baxter (University of New South Wales, Australia); Hamed Nosrati (CSIRO, Australia); Elias Aboutanios (University of New South Wales, Australia)
Increasing pressure on the available spectrum, particularly from wireless communications services, has led to significant research into strategies for coexistence between radar and communications. Dual function radar communications (DFRC) systems have emerged in this context as a potential solution to the spectrum congestion problem. DFRC schemes treat the radar as the primary modality and aim to embed the communications symbols into the transmitted waveforms. One such scheme is to use the frequency hops in a Frequency-Hopped (FH) MIMO radar to carry the communications information. While this scheme allows for higher data rates, embedding the information in the fast time impacts the radar performance. Therefore, the code book selection is an important design consideration that we focus on in this work. Specifically we consider two choices of the subset of codes and study their effects on the radar and communications performance. This work is an important step towards the goal of deriving the optimal code selection.
4:00 Optimum Code Design Using Genetic Algorithm in Frequency Hopping Dual Function MIMO Radar Communication Systems
Indu Priya Eedara, Moeness G. Amin and Ahmad Hoorfar (Villanova University, USA)
We consider a code-shift keying (CSK) dual-function radar-communication (DFRC) system where the radar implements frequency hopping wideband signal that is modulated by the communication symbols. Each symbol is represented by a phase coded waveform with derivative phase-shift keying (DPSK) modulation. The information sequences are multiplied with the FH waveforms in fast-time and are transmitted through the multiple-input multiple-output (MIMO) radar platforms. We use the genetic algorithm (GA) to design the optimum CSK sequences to reduce the radar range sidelobes modulation (RSM) which is imperative for clutter mitigation. It is shown that the optimal CSK sequences yield a substantial reduction in the range sidelobe level (RSL) fluctuations by simultaneously reducing the DFRC system's RSLs. Additionally, the proposed DFRC system with the optimized waveform design provides good spectrum containment and achieves high communication data rates by enabling the repeated use of hopping code values in the FH code design.
4:20 Phase Modulated Communications Embedded in Correlated FH-MIMO Radar Waveforms
Xiangrong Wang (Beihang University, China); Aboulnasr Hassanien (Wright State University, USA)
Methods for embedding communication symbols into the emission of MIMO radar with orthogonal frequency hopping (FH) waveforms have been recently proposed. These methods require that the FH waveforms be orthogonal within each sub-pulse to enable symbol detection at the communication receiver. This orthogonality constraint puts a limit on the total number of communication symbols that can be embedded within each pulse. In this paper, we propose a new signaling strategy for embedding communication symbols into correlated FH-MIMO radar waveforms. It is shown that by relaxing the orthogonality constraint, either the FH code length or antenna number or both can be increased leading to higher data rates. We also propose a novel technique for communication symbol detection based on frame representation. To reduce the computational load associated with the proposed detector, a fast detection algorithm is presented. A method for optimal FH code design is formulated by jointly optimizing the radar and communication operations. The ambiguity function and spectral efficiency are utilized to examine the effectiveness of the proposed strategy. Simulation results demonstrate that the symbol error rate (SER) can be significantly reduced by optimally designing frequency hops among different antennas with improved spectrum efficiency.

Wednesday, September 23 4:40 - 5:00 (Europe/Rome)

Coffee Break

Wednesday, September 23 5:00 - 5:20 (Europe/Rome)

WA-D1: DEMO 1

Room: Ch1
5:00 Reconfigurable, Low-Cost Monostatic Ultra-Wide Band Radar
Maxime Schutz (University of Limoges, XLIM & INOVEOS SAS Company, France); Cyril Decroze (XLIM, France); Michele Lalande (University of Limoges, France); Bertrand Lenoir (INOVEOS SAS, France)
An original monostatic UWB FMCW radar architecture has been developed for GPR applications. The radiation system is composed of a single antenna associated with a double switching configuration to minimize the compactness of the antenna system while ensuring a sufficient dynamic range. Moreover, an analog correction technique has been developed to linearize a low-cost free-running UWB Voltage Controlled Oscillator (VCO) and to enable operating frequency selection. A functional demonstrator has been designed and associated measurements are proposed.

WA-D2: DEMO 2

Room: Ch2
5:00 Demonstration of High-Fidelity Modeling & Simulation for Cognitive Radar
Sandeep Gogineni (ISL, Inc., USA); Joseph R. Guerci (Information Systems Laboratories, Inc. USA, USA); Hoan Nguyen (Information Systems Laboratories, USA); Jamie Bergin (ISL, USA); Brian Watson (ISL, Inc., USA); Muralidhar Rangaswamy (AFRL, USA)
Cognitive radar has emerged as key enabling technology to meet the demands of ever increasingly complex and congested radio frequency (RF) operating environments. The generally non-stationary, heterogeneous, and time-varying nature of the modern RF environment all but precludes the use of traditional adaptive processing methods that require the existence of wide sense stationary (WSS) training data. This necessitates an advanced modeling & simulation (M&S) framework that captures much of the realworld physics that gives rise to the aforementioned challenges such as heterogenous clutter, dense background targets, and intentional/unintentional radio frequency interference (RFI). As part of this demonstration, several examples that capture all these effects will be presented using the high-fidelity RF M&S tool RFView.

WA-D3: DEMO 3

Room: Ch3
5:00 A Hardware Prototype of Sub-Nyquist Radar with Distorted Pulse Shape
Satish Mulleti and Yariv Shavit (Weizmann Institute of Science, Israel); Moshe Namer (Technion - Israel Institute of Technology, Israel); Yonina C. Eldar (Weizmann Institute of Science, Israel)
Sub-Nyquist radar systems operate at lower sampling rates compared to the Nyquist rate and hence reduce the hardware cost and complexity. Sub-Nyquist systems uses the knowledge of the transmit pulse and the receive signal model to estimate the targets from lowrate samples. However, in practice, the pulse shape is often distorted and unknown at the receiver. Recently, a multiple-receiver based sub-Nyquist radar was proposed that estimates the targets without knowledge of the pulse shape. In this demo, we build a hardware prototype to demonstrate the proposed sub-Nyquist radar. We show that while operating at 10 times below the Nyquist rate the proposed two-receiver system with unknown pulse has comparable performance to a single-receiver sub-Nyquist system with a known pulse.

Wednesday, September 23 5:20 - 7:20 (Europe/Rome)

WA-P1: PSS2 - Advanced radar waveform design strategies

Chairs: Mohammad Alaee-Kerahroodi (Interdisciplinary Center for Security, Reliability and Trust, Université du Luxembourg, Luxembourg), Augusto Aubry (Universita degli studi di Napoli, Italy)
Range-Doppler Decoupling and Interference Mitigation Using Cognitive Random Sparse Stepped Frequency Radar
Kumar Vijay Mishra (United States Army Research Laboratory, USA); Satish Mulleti and Yonina C. Eldar (Weizmann Institute of Science, Israel)
Stepped frequency waveform (SFW) radars are used for synthesizing high range resolution profiles (HRRP). SFW radars suffer from strong range-Doppler coupling and are not robust to coexisting spectral interference. In this paper, we propose a new random, sparse step-frequency radar (RaSSteR) waveform to address these shortcomings. Unlike SFW where the carrier frequency is linearly increased over the available bandwidth, RaSSteR randomizes the frequency sequence to decouple range and Doppler. This new waveform also skips portions of the transmit spectrum without decreasing the range resolution and operates cognitively by focusing all its power in only a few frequencies. We derive theoretical guarantees which demonstrate that, even with few subpulses, RaSSteR has identical target recovery performance as the conventional random stepped frequency (RSF) waveform. Numerical experiments show that RaSSteR's target hit rate has a 30% improvement over the conventional RSF.
Transmit Signal and Receive Filter Design of PMCW Radar with Low-Resolution ADC
Foozie Foroozmehr (Isfahan University of Technology, Iran); Mohammad Mahdi Naghsh (Isfahan University of Technology & Uppsala University, Iran); Mahmood Modarres-Hashemi (Isfahan University of Technology, Iran)
In this paper, we consider the problem of transmit signal and receive filter design for a phase modulated continuous wave (PMCW) monostatic single-input single-output (SISO) radar system when low-resolution analog-to-digital-convertor (ADC) is used at the receiver. By applying Bussgang decomposition model, the effect of low-resolution ADC is considered in the model, and two cases of colored and white noise are investigated. In the white noise case, it is concluded that the matched filter (MF) is the best filter maximizing the signal-to-interference-plus-noise-ratio (SINR) of the receive filter output. In the presence of colored noise, we propose a cyclic design procedure for maximizing the output SINR. Numerical examples illustrate the superiority of the proposed method in comparison with the well-known m- sequence probing signal with matched or designed receive filter. The effect of different parameters is also verified.
Finite Alphabet Unimodular Sequence Design with Low WISL via an Inexact ADPM Framework
Xianxiang Yu (University Of Electronic Science And Technology Of China, China); Guolong Cui (University of Electronic Science and Technology of China (UESTC), China); Jing Yang and Lingjiang Kong (University of Electronic Science and Technology of China, China)
This paper studies the unimodular sequence design with the finite alphabet case for a cognitive radar in order to achieve the desired auto-correlation. An iteration algorithm based on Inexact Alternating Direction Penalty Method (IADPM) framework is developed to minimize the Weighted Integrated Sidelobe Level (WISL). In each iteration, it splits the NPhard problem into two subproblems via an introduced auxiliary variable, while locally increasing the penalty factor involved in the IADPM framework. The proposed algorithm is shown to converge for any initialization under some mild conditions and avoids the non-convergence problem of ADMM when handling the NP-hard problems. Finally, the auto-correlation performance and convergence speed of the proposed algorithm are evaluated against the state-of-the-art methods. Results show that our proposal outperforms the state-of-the-art competing methods while providing the favorable performance-complexity balance.
ADMM Based Transmit Waveform and Receive Filter Design in Cognitive Radar Systems
Ehsan Raei (SnT - Interdisciplinary Centre for Security, Reliability and Trust, University of Luxembourg, Luxembourg); Mohammad Alaee-Kerahroodi (Interdisciplinary Center for Security, Reliability and Trust, Université du Luxembourg, Luxembourg); Bhavani Shankar Mysore R (Interdisciplinary Centre for Security, Reliability and Trust & University of Luxembourg, Luxembourg)
In this paper, we propose an attractive method to jointly design discrete phase radar sequence and receive Doppler filter with the aim of enhancing Signal to Interference and Noise Ratio (SINR) in a cognitive radar system. Towards this, we consider minimizing the output power of interference under Capon constraint, when transmitting $M$-ary Phase Shift Keying (MPSK) sequences. This minimization results in a multi-variable and non-convex optimization problem. To tackle the problem we propose an efficient algorithm based on the Alternating Direction Method of Multipliers (ADMM) to optimize the Doppler filter and waveform alternatively. The problem with respect to filter is convex enabling a closed-form solution, while the problem is non-convex with regards to the waveform. We propose a Coordinate Descent (CD) framework to address this aspect and illustrate some numerical results.
Diversifying the Processing Chain of the Modified Stepped Frequency Radar Waveform Combined with Pulse Compression Techniques
Mahdi Saleh (Lebanese University, Lebanon); Samir Omar (Lebanese International University, Lebanon); Eric J. Grivel (Université de Bordeaux, France)
A processing chain from the transmitter to the receiver dealing with a stepped frequency waveform combined with phase coding (PC) has been proposed in a recent paper. It consists in splitting the spectrum of a PC pulse into predetermined portions, and then transmitting the corresponding time-domain signals. At the receiver, a modified version of the frequency domain (FD) algorithm is used to process the received echoes. However, this processing chain could be improved. Indeed, alternative modulations could be considered to offer a degree of freedom to the practitioner. In addition, when the number of portions becomes large, the performance of the proposed FD algorithm degrades, the spectrum of the reconstructed waveform being distorted. Therefore, in this paper, the novelty stands in the following items 1) at the transmitter, the non-linear frequency modulation (NLFM) can be considered as an alternative to PC 2) at the receiver, a novel time-domain waveform reconstruction (TWR) is proposed to process the received echoes in the time domain instead of operating in the frequency domain. Our simulations show that, unlike the modified FD algorithm, for any number of portions, the performance obtained using the TWR algorithm is the same at high signal-to-noise ratio (SNR), whereas at low SNR, the larger the number of portions used, the better the performance.

WA-P2: Localization and estimation

Chairs: Francesco Bandiera (University of Salento, Italy), Shenghua Zhou (National Laboratory of Radar Signal Processing, Xidian University, China)
Angular Target Detection by Means of DPCA like Technique
Enzo Carpentieri (Mbda Italy, Italy); Gregorio De Stefano and Luca Musetta (MBDA, Italy); Giovanni Petraglia and Pietro Vinetti (MBDA Missile Systems, Italy)
Displacement Phase Center Antenna (DPCA) technique is largely adopted in clutter suppression from moving platforms. In this paper, using a DPCA technique, it has been evaluated the capability of discriminating the angular position of targets, by means of the induced Doppler obtained by platform motion emulation. Using a linear array and sequentially activating the antenna elements, it has been determined the targets positions from the echo signal collected at each antenna port. An experimental setup for has been developed in order to validate the approach.
Joint Range and Doppler Estimation with Amplitude Weighted Linearly Constrained Minimum Variance Method for OFDM-Based RadCom Systems
Fuqiang Zhang, Zenghui Zhang and Wenxian Yu (Shanghai Jiao Tong University, China)
Recently, the joint operation of radar and communication (RadCom) systems becomes popular and an orthogonal frequency division multiplexing (OFDM) signal can be used to realize this system. In order for an OFDM-based RadCom system with high performance, accurate range and Doppler estimation needs to be achieved. For this purpose, the DFT method, which can be implemented very fast, is used commonly. However, this method greatly suffers from high sidelobes such that the weak target could get masked. The estimation results become even worse in the strong clutter case. To deal with this problem, an amplitude weighted linearly constrained minimum variance (AW-LCMV) method is proposed. In our proposed AW-LCMV method, the problem of high sidelobes is well handled by fully exploiting the target information in the range-Doppler domain and adding a regularization term on the range dimension. In turn, much better estimation results can be obtained by the AW-LCMV method. Finally, excellent simulation results are used to demonstrate the effectiveness of our proposed method.
An Improved Closed-Form Solution for Differential RSS-based Localization
Yimao Sun (China Academy of Engineering Physics & University of Electronic Science and Technology of China, China); LiXianglu (iee of CAEP, China); Zhijiang Huang and Jie Tian (China Academy of Engineering Physics, China)
One of the advantages of differential received signal strength (DRSS)-based localization is eliminating the dependence of the source transmit signal power. Therefore, DRSS is able to handle the passive localization as time difference of arrival (TDOA) does. The recent work of [1] investigated a new model for DRSS measurements, but the proposed methods are suboptimal. This paper focuses on the study of localization using DRSS measurements. A new formulation for DRSS localization of a 2-D source is proposed, which leads to a closed-form estimator through minimizing a constrained optimization problem. The solution applies weighted least squares technology only, achieving the Cram\'er Rao Bound (CRB) accuracy if the noise is mild. A brief analysis validates that the mean-square error (MSE) attains the CRB in the small noise region. The performance and theoretical result are verified by simulations.
Range-Doppler Spectrum Estimation Based on Matrix Completion for Frequency Agile Radar
Xueyao Hu (Beijing Institute of Technology Chongqing Innovation Center, China); Fugang Lu (No. 203 Research Institute of China Ordnance Industries, China); Can Liang (Beijing Institute of Technology, China); Jianhu Liu (Beijing Rxbit Electronic Technology Co., Ltd, China); Yanhua Wang (Beijing Institute of Technology, China)
Coherent frequency agile (FA) radar has attracted considerable attention. Compared with the stepped frequency radar, FA radar not only has the high range-resolution characteristic obtained by synthesizing large bandwidth, but also has the ability of the range-Doppler decoupling and better electronic countermeasures performance. Typically, traditional matched filter (MF) is often used to estimate the range-Doppler spectrum for FA radar. However, the processing results of the MF method are affected by the raised sidelobe pedestal, which leads to the degradation of radar performance in target detection and parameter estimation. In this paper, we propose a matrix completion (MC)-based method for range-Doppler spectrum estimation. Based on the concept of the FA pattern, the echo is modeled as a time-frequency observation matrix, and we use the MC technique to jointly recover the unobserved data in the matrix to estimate the range-Doppler spectrum with sidelobe pedestal suppression. Moreover, compared with the FA radar with one-dimensional compressed sensing recovery, the MC-based method overcomes the grid mismatch problem, further enhancing the range-Doppler spectrum estimation performance. The simulation results verify the effectiveness of the proposed method.
A-4 Skyhawk Aircraft Stealth Capacity Against L-Band Radar Based on Dynamic Target Detection
Renan Richter (Instituto Tecnológico de Aeronáutica, Brazil); Newton Gomes (Instituto Tecnológico de Aeronáutica - ITA, Brazil)
The purpose of the paper is to analyze the A-4 Skyhawk aircraft stealth capacity as a function of its radar cross section (RCS) against an L-band monostatic radar (1 to 2 GHz). The article proposes a trajectory to a certain probability of detection, a statistical variable, by reaching a 81.3% reduction of the theoretical radar range. This methodology evaluates the effect of dynamic target detection when the aircraft performs several flight paths towards a radar site region, looking for an altitude that minimizes the radar range and, consequently, increasing A-4 crew's combat survivability.
Radon-Fourier Transform in FMCW Radar
Longhui Wang and Jian Wang (Tsinghua University, China)
This paper studies the long time coherent integration (LTCI) algorithm in frequency modulation continuous wave (FMCW) radar for weak maneuvering target detection. A new coherent integration algorithm, called FMCW Radon-Fourier transform (FMCW-RFT), is proposed. This algorithm eliminates range migration (RM) via two-dimensional searching in range-velocity parameter space, removes range-velocity coupling (RVC) in FMCW radar via phase compensation function (PCF) and achieves coherent integration. The addressing operation and the phase compensation operation of FMCW-RFT are different from the existing Radon-Fourier transform based on pulse Doppler (PD-RFT) radar due to different radar systems. Finally, some simulations confirm the effectiveness of FMCW-RFT. The results show that FMCW-RFT performs better than PD-RFT and two-dimensional fast Fourier transform (2D-FFT) on integration gain and detection ability in low signal-to-noise ratio (SNR) scenario.
Searching Velocity Fourier Transform: A Novel Focus-Before-Detection Algorithm in FMCW Radar
Longhui Wang and Jian Wang (Tsinghua University, China)
In this paper, the focus-before-detection (FBD) algorithm in frequency modulation continuous wave (FMCW) radar is considered and a novel FBD algorithm for weak maneuvering target detection, called searching velocity Fourier transform (SVFT), is proposed. This algorithm eliminates the fast-time and the slow-time coupling (FSC) caused by target motion via searching in the velocity parameter space and improves the focus performance of FMCW radar. First of all, SVFT compensates beat signal and removes FSC by searching velocity function (SVF). Secondly, the compensated beat signal is focused by twodimensional fast Fourier transform (2D-FFT) in the candidate Doppler bins (CDB) which are related to SVF, and the target energy can be integrated into a peak. Then, the peak is obtained by constant false alarm rate (CFAR) detection. After that, according to the result of CFAR detection, the accurate range and velocity of the target can be obtained by solving equations. Finally, some simulations confirm the effectiveness of SVFT. The results show that SVFT performs better on focus ability and detection ability with lower computational complexity.
Low-Rank Approximation-Based Super-Resolution Imaging for Airborne Forward-Looking Radar
Jie Li (University of Electronic Science and Technology, China); Yongchao Zhang, Yin Zhang and Yulin Huang (University of Electronic Science and Technology of China, China); Jianyu Yang (School of Electronic Engineering, China)
Iterative adaptive approach (IAA), based on the weighted least squares estimation (WLS) criterion, can effectively improve the azimuth resolution of airborne forward-looking radar imagery. Regretfully, the brute force IAA requires a large number of inversions of high-dimensional autocorrelation matrix, resulting in notably high computational complexity. In this paper, a low-rank iterative adaptive approach (LR-IAA) is proposed to solve this problem. Our underlying idea is to construct a low-rank Gaussian matrix to randomly sample the original echo model, and restore the original scene through spectral estimation method in a low-dimensional linear space. Compared with brute-force implementation, the proposed LR-IAA enjoys a preferable computational efficiency without performance degradation. Simulations are given to verify the performance gain.
Radar SLAM for Autonomous Indoor Grinding
Nils Mandischer, Sami Charaf Eddine, Mathias Huesing and Burkhard Corves (RWTH Aachen University, Germany)
In the EU project Bots2Rec an autonomous asbestos reconstruction unit is developed by the Institute of Mechanism Theory, Machine Dynamics and Robotics (IGMR) of RWTH Aachen University. In the hazardous environment the mobile robot is expected to localize itself relative to the walls. Due to the heavy dust formation during and after grinding tasks, solely laser-based sensing is not reliable. Therefore, Bot2ReC deploys an additional 2D radar sensor to compensate for restricted visibility. In this work the developed online radar SLAM for indoor environments is presented. This method is based on the probabilistic iterative correspondence algorithm utilizing the Mahalanobis metric and improves it towards online capability. Further, a radar filter is deployed to reduce the data load.
Polarization Parameter Estimation of Conformal MIMO Radar Targets
Xin Wang (Xidian University, China); Shenghua Zhou (National Laboratory of Radar Signal Processing, Xidian University, China); Hongwei Liu (National Laboratory of Radar Signal Processing, China); Junkun Yan (Xidian University, China); Hongtao Su (National Laboratory of Radar Signal Processing,China); Huijing Li (Xidian University, China); Kuiying Yin (Nanjing Research Institute of Electronic Technology, China); Wei Sun (Nanjing Electronic Technology Research Institute, China); Lin Zhou (Southeast University, China)
Conformal Multiple-Input Multiple-Output (MIMO) radar make good use of the aperture of the platform, but target polarization scattering parameter is more complicated since the beam patterns of transmit/receive antennas are different from those in free space. In this paper, we develop signal processing algorithms to estimate the Sinclair matrix of radar targets for conformal MIMO radar, according to the Least Square (LS) criterion and the Maximum Likelihood (ML) criterion. Both unbiased but the ML estimator can achieve the Cramer-Rao bound. The estimation performance is evaluated in theory via numerical results.
Parameter Estimation and Imaging of Three-Dimensional Moving Target in Dual-Channel CSAR-GMTI Processing
Beibei Ge, Daoxiang An and Zhimin Zhou (National University of Defense Technology, China)
Circular synthetic aperture radar (CSAR), with high resolution and multi-angular apertures, is an innovative mode which can perform consecutive surveillance and target tracking. The important purpose of moving target tracking is parameter estimation. In this paper, we present a method to estimate the motion parameters and refocus imaging of three-dimensional (3D) moving target. Doppler information and interferometric phase term are the basic of motion parameter. And the estimated results serve as the prior knowledge of refocus imaging. Simulation results demonstrate the correctness of the analysis and the performance of the proposed approach.
Fixed Budget Kernel LMS Based Estimator Using Random Fourier Features
Aditya Ramesh (IIIT Sricity, India); Uday Singh (Indian Institute of Technology Indore, India); Rangeet Mitra (ETS Montreal, Canada); Vimal Bhatia (Indian Institute of Technology Indore, India); Amit Mishra (University of Cape Town, South Africa)
Accurate estimation of delay and Doppler shift are essential for target detection and tracking in a radar system. In this regard, online reproducing kernel Hilbert space (RKHS) based estimation techniques have emerged as viable for radar systems, due to guarantees of universal representation, and convergence to low estimator variance. However, existing RKHS based estimation techniques for radar rely on growing dictionary of observations, which makes it difficult to predict the memory requirement beforehand. Furthermore, online dictionary based learning techniques are prone to erroneous observations in the high-noise regime. In this work, a finite implementation-budget estimator is proposed, which utilizes an explicit mapping to RKHS using random Fourier features (RFF). The proposed RFF based estimator achieves equivalent/better performance as compared to its dictionary-based counterpart and has a finite memory requirement that makes the estimator attractive for practical implementation. Simulations are performed over realistic radar scenarios, that suggest the viability of the proposed RFF based estimator.

WA-P3: Waveform diversity and MIMO radars

Chairs: Enzo Carpentieri (Mbda Italy, Italy), Giuseppe Ricci (University of Salento, Lecce, Italy)
Unified Metric-Based Multi-Platform Network: Integrating Radar, Communication and Jamming
Yuanhan Ni and Zulin Wang (Beihang University, China); Qin Huang (Group 201, Beihang University, China)
Besides radar and communication, this paper investigates the integration of three functionalities including jamming. To this end, a unified metric-based network system consisting of multi-platform with different capabilities is proposed to integrate radar, communication and jamming functionalities. In such a system, radar target, communication user and jammed target are regarded as the virtual generalized users. To facilitate the collaboration of multi-platform, all generalized users have the designed unified signal-to-noise ratio (SNR) metric. Given the SNR model, two optimization-based transmit beamforming methods are designed: 1) minimizing the total power and 2) maximizing the SNR at anyone radar target. The designed beamforming methods jointly precode radar, communication and jamming waveforms simultaneously, which are non-convex, but can be solved by the classic semidefinite relaxation (SDR) technique. Numerical results show that the proposed multi-function multi-platform (MFMP) network can significantly promote the overall performance of integrated system compared with the single-platform system.
Adaptive Transmit Power Allocation for FDA Radar with Spectral Interference Avoidance
Ronghua Gui and Wen-Qin Wang (University of Electronic Science and Technology of China, China)
In this paper, we first establish a spectral interference model for frequency diverse array (FDA) radar, with a focus on the interference covariance matrix structure. Then, we show that the FDA radar has a capability in suppressing spectral interferences, even if they enter through the receiver mainlobe. In order to improve the suppression performance, we further design a transmit weight vector for adaptive element-wise power allocation, through maximizing the output signal-to-interference-plus-noise ratio (SINR). Numerical results demonstrate that the FDA radar with adaptive power allocation is able to avoid those spectral interferences via adjusting its transmit spectrum. The resulting output SINR is higher than those of the FDA with uniform power allocation, conventional phased-array (PA) and multi-input multi-output (MIMO) radars.
Design of (Quasi) Complementary Waveform with Doppler Resilience for Range Sidelobe Suppression
Zhongjie Wu (Harbin Institute of Technology, China); Chenxu Wang (Harbin Institute of Technology,Weihai, China); Zhiquan Zhou (Harbin Institute of Technology, Weihai, China); Xiang Feng (Harbin Institute of Technology, China)
This paper is devoted to constructing Doppler resilient (quasi) complementary waveforms (DRCW/DRQCs) out of conventional complementary pairs via the joint design of both transmit pulse trains and receive pulse weights. The design of DRCWs is formulated to maximize the SNR under several Doppler null point constraints. It turns out to be a two-way partitioning problem, and is approximately solved by semidefinite programming relaxation and randomization techniques. The design of DRQCWs is a spectrum shaping problem with a low modulus variation constraint. A suboptimal solution can be efficiently computed via FFT under the Majorization-Minimization framework. Our methods can generate DRCW/DRQCWs with multiple range sidelobe blanking areas at arbitrary Doppler shifts and acceptable decreases in the SNR and Doppler resolution. Based on those properties, a novel waveform emission scheme for cognitive radar is proposed, which shows great promise in many applications like detecting small-RCS and low-speed targets.
Waveform Design for MIMO Radar with Partial Low-Resolution ADCs
Ziyang Cheng (University of Electronic Science and Technology of China, China); Jinyang He (UESTC, China); Bin Liao (Shenzhen University, China); Zishu He (University of Electronic Science and Technology of China, China)
This paper proposes an optimization scheme for collocated multiple-input-multiple-output (MIMO) radar with a partial low-resolution ADCs architecture. With a criterion of the signal-to-interference plus quantization noise ratio (SIQNR) maximization, we formulate our problem by jointly optimizing the waveform, ADC switch vector and receive filter. Since the resultant problem is nonconvex, an alternating optimization framework (AOF) is proposed. Specifically, the block successive upper-bound minimization (BSUM) framework based on the Dinkelback method is devised to optimize the ADC switch vector and waveform matrix. Moreover, we propose a continuous and concave function to approximate the \(l_0\) norm to obtain a high-quality sparse solution to the ADC switch vector. Numerical simulations are provided to demonstrate the effectiveness of the proposed schemes.
A Mismatched Filter for Integrated Sidelobe Level Minimization over a Continuous Doppler Shift Interval
Maria-Elisavet Chatzitheodoridi (ONERA, France); Abigael Taylor (ONERA, the French Aerospace Lab, France); Olivier Rabaste (Onera, France)
A mismatched filter that minimizes the phase code sidelobe level in the presence of Doppler shifts is presented. Taking into account the Doppler shift is essential, as for non-robust to Doppler shift phase codes, the sidelobe level may increase dramatically. The proposed filter is obtained as the solution of an optimization problem that is based on the use of the mean Integrated Sidelobe level over a given Doppler interval. Its cost function can be simply expressed as the product between the zero-Doppler cost function and the generating matrix of the Discrete Prolate Spheroidal Sequences. Comparison with known filters from the literature, for different filter lengths and Doppler interval ranges, shows that the proposed solution provides improved results in terms of mean Integrated Sidelobe Ratio over the considered Doppler interval.
CP-OFDM Radar Range Reconstruction in a Chaotic Doppler Disturbed Scenario
Xinyu Liu (University of Electronic Science and Technology of China & School of Communication and Information Engineering, China); Tianxian Zhang and Qiao Shi (University of Electronic Science and Technology of China, China); Lingjiang Kong (University of Electronic Science and Technology of China (UESTC), China)
Considering the serious performance degradation of cyclic prefix based orthogonal frequency division multiplexing (CP-OFDM) radar range reconstruction caused by chaotic Doppler disturbance, an inter-carrier interference (ICI) cancellation problem is investigated in the scene of multiple unknown moving targets. The ICI cancellation problem is solved by proposing a two-step subcarrier alignment method. Firstly, we construct the signal model and separate the intra-pulse and inter-pulse Doppler modulation term. Then, the two-step orthogonal alignment of the subcarrier method is put forward to against ICI by Fourier analysis on the converted signal model. Later, inter-range-and-carrier (IRCI) free range reconstruction is implemented with the assistance of the CP-OFDM pulse compression method. Finally, numerical results are provided to verify the validity of the proposed algorithm.
Ergodic Interference Steering for Joint Phased Array Radar and Communication Systems
Bingqing Hong and Wen-Qin Wang (University of Electronic Science and Technology of China, China)
In this paper, we propose an ergodic interference steering (IS) method for co-existing radar-communication systems that consists of a phased array (PA) radar and multi-user multiple-input multiple-output (MIMO) communication systems. Although ergodic interference alignment (IA) method have been proved the mutual interferences between MIMO radar and MIMO communications can be efficiently eliminated, it will increase transmitter power consumption. Differently, our proposed ergodic IS allows to reduce the cost of total transmit power. In addition, we exploit the proposed integrated architecture to ensure efficient radar detection performance. The proposed method is evaluated by detection probability for radar functionality, while the communication functionality is examined by bit error rate (BER). Both theoretical analysis and simulation results validate that the effectiveness of the ergodic IS method for co-existing radar and communication systems.
Fast Optimization for Unimodular Sequences Design with Good Correlation Properties
Yi Bu and Jing Yang (University of Electronic Science and Technology of China, China); Xianxiang Yu (University Of Electronic Science And Technology Of China, China); Yanqin Xu (University of Electronic Science and Technology of China, China); Guolong Cui (University of Electronic Science and Technology of China (UESTC), China); Zhihao Jiang (Naval Research Academy, China)
This paper deals with unimodular sequences design with good correlation properties based on a fast Iterative Sequential Quartic Optimization(ISQO) algorithm. The Weighted Integrated Sidelobe Level (WISL) of auto- and cross-correlation is minimized as a figure of metric. In each iteration, the nonconvex quartic optimization problem is decomposed into a quadratic problem which can be converted to a linear optimization problem with closed-form solutions. Herein, a special algebraic structure of the objective function is explored. Finally, numerical examples highlight that the proposed algorithm shows a faster convergence speed while guaranteeing that the sequences with low correlation level.
Correlation-Based Radar Receivers with Pulse-Shaped OFDM Signals
Steven Mercier (ISAE-SUPAERO, University of Toulouse, France); Damien Roque (ISAE-SUPAERO, Université de Toulouse, France); Stéphanie Bidon (University of Toulouse / ISAE, France); Cyrille Enderli (Thales Airborne Systems, France)
In waveform sharing scenarios, various radar receivers have been developed for orthogonal frequency-division multiplexing (OFDM) signals. More general waveforms, such as pulse-shaped multicarrier modulations received little attention so far, despite their increased robustness to high-Doppler scatterers. In this paper, we compare the performance of two correlation-based radar receivers, namely the matched filter and the symbol-based technique, when used with different pulse-shaped multicarrier waveforms. We express the signal-to-interference-plus-noise-ratio in the range-Doppler map, taking into account the pedestal (or random sidelobes) induced by the symbols. Benefits of pulse shaping is further illustrated in a realistic vehicular scenario, in presence of multiple targets and ground clutter. In this context, the symbol-based approach outperforms the matched filter while enjoying a low-computational complexity. More generally, our results reveal the multicarrier pulse shape as a relevant degree of freedom in waveform co-design approaches (\emph{e.g.}, cognitive radar/communication systems).
FPGA Design and Implementation of a Real-Time FM/PM Pseudo Random Waveform Generation for Noise Radars
Alisson Barreto (Instituto Militar de Engenharia, Brazil); Leandro Guimarães Figueroa Pralon (CTEX, Brazil); Bruno Pompeo and Mariana G Pralon (Brazilian Army Technological Center, Brazil); Gabriel Beltrão (University of Luxembourg & Interdisciplinary Centre for Security, Reliability and Trust, Luxembourg)
Noise Radar technology is the general term used to describe radar systems that employ realizations of a given stochastic process as transmit waveforms. With the advances made in hardware as well as the rise of the software defined noise radar concept, waveform design emerges as an important research area related to such systems. Several optimization algorithms have been proposed to generate pseudo-random waveforms with specific desired features, specially with respect to side lobes. Nevertheless, not only modifying random waveforms may compromise their LPI performance, but also the implementation of such algorithms in real time applications may not be feasible. Within this context, this paper analyzes varied design architectures for FM/PM pseudo-noise waveform generation, considering a real-time application. The proposed architectures are verified in a co-simulation environment using Xilinx R○ System Generator tool and implemented on reconfigurable hardware, i.e., a Xilinx Field Programmable Gate Array (FPGA) is taken into consideration. Timing, resources consumption and the trade offs related to hardware area and performance are then investigated.
Fractional Wideband Ambiguity Function for MIMO Radar
Chang Gao (Beijing Institute of Technology, China); Yongzhe Li (Beijing Institute of Technology & School of Information and Electronics, China); Ran Tao (Beijing Institute of Technology, China)
We define a new ambiguity function (AF) which involves fractional signal processing and meanwhile incorporates all functions of conventional AFs for the multiple-input multiple-output (MIMO) radar with wideband signals. We term the newly defined AF as the fractional wideband MIMO radar AF, for which we provide explicit derivations and also present explanations and indications. To obtain this, we start from introducing a factor that relates to the fractional signal processing into the matched-filtering exploited by the AF definition, in order to provide a new perspective on evaluating the ambiguity property. We then derive the fractional wideband MIMO radar AF into a simplified version for the case of slow moving targets, and also develop its properties. Moreover, we establish relationships between the newly defined AF and existing ones, and prove that it can serve as a generalized AF form. Simulation results show that the fractional wideband AF gives lower sidelobe levels when the fractional order is properly adopted.
Direction Blanking with Controlable Width in MIMO Radar via Phase Code Optimization
Olivier Rabaste (Onera, France)
In this article, we consider the problem of creating null-energy directions of controlable width with a MIMO radar, for surveillance applications where the energy radiated should be uniformly spread over the remaining angular directions. In a MIMO radar, the antenna beam pattern can be modified by a judicious choice of the transmitted waveforms. We restrict our attention here to phase code waveforms. We first propose an optimization program that enables to minimize the energy in a given narrow direction while uniformly spreading the energy over the other directions. Then we propose to extend this solution to create larger null angular intervals. Interestingly the proposed solution involves the well-known Discrete Prolate Spheroidal Sequences. We show on simulations the efficiency of the proposed approach.
A Learning Approach to Design Binary Sequences with Good Correlation Properties
Omid Rezaei and Mahdi Ahmadi (Isfahan University of Technology, Iran); Mohammad Mahdi Naghsh (Isfahan University of Technology & Uppsala University, Iran)
Employing binary sequences with good auto- correlation properties leads to performance improvement for active sensing systems. Designing such sequences has been studied widely in the literature from optimization point of view. In this paper we consider the problem of designing binary sequences with good aperiodic/periodic auto-correlation functions in terms of Integrated Sidelobe Level (ISL) using machine learning method. Specifically, we propose a novel neural network structure which can learn an algorithm for designing binary sequences with small ISL. We also extend the resulting network, which we refer to as BiSCorN, by designing Low-Correlation Zone (LCZ) binary sequences. Numerical experiments show that our proposed method outperforms state-of-the-art algorithms in terms of ISL, and interestingly Peak Sidelobe Level (PSL).
Doppler-Tolerant NLFM Radar Signal Synthesis Method
Hubert Milczarek, Czesław Leśnik and Adam Kawalec (Military University of Technology, Poland)
The problem of designing a radar NLFM waveform with reduced Doppler sensitivity is examined in this paper. A presented novel approach utilizes a U-shaped instantaneous phase functions to determine a signal time-frequency structure combined with optimization framework aimed to reach compromise between sidelobe attenuation and range resolution with additional requirement on Doppler tolerance. Spectral efficiency of the signal is also examined and accounted for. This aspect has been rarely discussed to date, whereas it needs to be considered in practical implementation due to the limits arising from band allocation and interference mitigation point. The simulation results show that the signals synthesized with proposed method reach sidelobes level as low as -50 dB without any amplitude modulation. Finally, improved Doppler tolerance makes proposed waveform attractive for surveillance radar.
Performance Prediction of FDA-MIMO Radar Detector
Ziting Xu, Bang Huang and Wen-Qin Wang (University of Electronic Science and Technology of China, China)
Due to the complex shape of the target, the radar cross-section (RCS) scintillations of the target will affect the radar detection performance. Therefore, a signal model for frequency diverse array (FDA) multiple-input and multiple-output (MIMO) radar based on the distributed source model is constructed in this paper, which takes advantage of spatial diversity. Furthermore, the detection performance of the FDA-MIMO radar multi-pulse detector is analyzed for targets conform to Swerling I and Swerling II model. Numerical results show its advantages over MIMO radar detector at low signal-to-noise ratio (SNR).

WA-P4: Automotive radar technology and signal processing

Chairs: Christos V. Ilioudis (University of Strathclyde, United Kingdom (Great Britain)), Alexander Yarovoy (TU Delft, The Netherlands)
Automotive Radar Interference Reduction Based on Sparse Bayesian Learning
Shengyi Chen (Ruhr-Universität Bochum & HELLA GmbH & Co. KGaA, Germany); Jalal Taghia (Ruhr-Universitaet Bochum, Germany); Uwe Kühnau, Tai Fei and Frank Grünhaupt (HELLA GmbH & Co. KGaA, Germany); Rainer Martin (Ruhr-University Bochum, Germany)
Automotive radar plays an important role in advanced driver assistant systems to support level-2 automated driving functions. However, the mutual interference between automotive radars increases due to the rising density of radars on the road. Therefore, the radar signal will be distorted to some extent and the performance of radars will degrade if no countermeasures are taken. In this paper, an interference mitigation approach using compressive sensing (CS) and Bayesian learning is introduced. By utilizing the sparsity of the beat signal in the frequency domain, the range-Doppler (RD) spectrum can be reconstructed with the help of undistorted samples in the beat signal. The sparse Bayesian learning method (SBL) is used to estimate the posterior of the signal's sparse representation and to infer the maximally sparse representation by using the Expectation-Maximization (EM) algorithm. It is shown that the SBL-based method has its advantages in signal-to-interference-plus-noise ratio (SINR) and target peak detection in comparison to conventional CS or classical signal reconstruction algorithms like linear predictive coding (LPC).
An Adaptive Interference Detection and Suppression Scheme Using Iterative Processing for Automotive FMCW Radars
Masahiro Umehira, Takeo Okuda, Xiaoyan Wang and Shigeki Takeda (Ibaraki University, Japan); Hiroshi Kuroda (Hitachi Automotive Systems, Ltd., Japan)
Millimeter-wave FMCW (Frequency Modulated Continuous Wave) radar is expected to be widely used for ADAS (Advanced Driver Assistance System) and automatic driving car. Dense deployment of FMCW radars causes serious inter-radar interference resulting in miss-detection and/or false detection of the targets. As there are various interference environments, adaptive inter-radar interference detection and suppression is required. This paper proposes an adaptive interference detection and suppression scheme based on iterative processing for automotive FMCW radars. The proposed scheme detects the wide band interference using adaptive interference detection threshold control and suppresses it using adaptive interference suppression windowing width control in order to improve SNR (Signal to Noise power Ratio). Computer simulation is conducted to validate the proposed scheme and the results confirm the proposed scheme can improve SNR in various interference environments.
A Ghost Target Suppression Technique Using Interference Replica for Automotive FMCW Radars
Daiki Ammen (Ibaraki University & Graduate School of Science and Engineering, Japan); Masahiro Umehira, Xiaoyan Wang and Shigeki Takeda (Ibaraki University, Japan); Hiroshi Kuroda (Hitachi Automotive Systems, Ltd., Japan)
As an automotive FMCW (Frequency Modulated Continuous Wave) radar used for automated driving or ADAS (Advanced Driver Assistance System) will be widely deployed, miss-detection and/or false detection due to inter-radar interference can be a serious problem in near future. In order to mitigate narrowband interference resulting in false detection or so-called ghost target detection, this paper proposes a novel ghost target suppression technique using interference replica for automotive FMCW radars where the narrowband interference replicas is generated by performing carrier sense before initiating radar signal transmission and suppresses the ghost target from the received beat signals by subtracting the generated interference replica from the received signals in the frequency spectrum. Computer simulations confirm the proposed ghost target suppression scheme is feasible.
A Graphical Heatmap Tool to Analyse the Effects of Interference in Automotive Radar
Fatemeh Norouzian (University of Birmingham, United Kingdom (Great Britain)); Anum Ahmed Pirkani (University of Birmingham, United Kingdom (Great Britain) & AKSA-SDS, Pakistan); Edward Hoare and Mikhail Cherniakov (University of Birmingham, United Kingdom (Great Britain)); Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain))
A unique graphical tool is presented in this paper that shows the severity of effects, in terms of signal to interference ratio, when receiving interference from various FMCW radars into one FMCW victim radar. The Heatmap tool is of importance as wide range of automotive radars are currently operating and more will be introduced in the future with plethora of new parameters. Understanding the effects on the functionality of the automotive radars in the presence of interference is crucial to aid development of mitigation strategies. The universal Heatmap tool can also be used to estimate the level of signal to interference ratio for any victim and interferer radars pair and road scenarios. The estimated signal to interference ratio value calculated by the described technique is compared and validated with simulation results and backed by laboratory experimental measurement results.
Designing Radar Waveform Parameters Oriented to Performance of Collision Warning System
Hang Ruan, Yimin Liu, Tianyao Huang and Xiqin Wang (Tsinghua University, China)
With the popularization of automotive radars, mutual interference among them becomes a noteworthy issue. One countermeasure is to improve the efficiency of radar's use of electromagnetic resources based on the performance of automotive system. In this paper, we take the collision warning system as an example and propose a method of designing the radar waveform parameters oriented by system performance. We propose total wrong decision loss as the metric to quantify the performance and study the relationship between this metric and radar waveform parameters (bandwidth and duration). Then, the waveform parameters are designed with a limit on the electromagnetic resources to optimize the system performance. Numerical results show that the proposed design outperforms the state-of-the-art parameter settings in terms of system performance and resource or energy efficiency.
Multiple Access Radar Using Slow Chirp Modulation
Hans Hellsten (Saab AB, Halmstad University); Emil Nilsson (Halmstad University, Sweden)
The cohabitation of several radars, operating in the same frequency band, has become an essential and urgent topic as active safety systems for automotive applications are rolled out. An obvious concern is that mutual interference must be managed. Separating users in time, i.e. TDMA, achieves the required level of isolation in a straightforward way. CDMA techniques providing sufficient channel isolation are less obvious. The paper develops an alternative CDMA method, called Slow Chirp Modulation (SCM). SCM utilizes the full coherent integration time for transmission of a single aperiodic but ergodic signal, allowing target range and velocity to be retrieved but minimizing spectral occupancy. Spectral efficiency two orders of magnitude higher than for the discussed alternative methods is obtained, allowing more than a thousand non-interfering channels. Relying on indicated hardware schematics, the paper demonstrates the functionality of the novel signal processing algorithms, which are required for SCM.
Statistical Analysis of Automotive Radar Interference
Anum Ahmed Pirkani (University of Birmingham, United Kingdom (Great Britain) & AKSA-SDS, Pakistan); Fatemeh Norouzian and Edward Hoare (University of Birmingham, United Kingdom (Great Britain)); Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain)); Mikhail Cherniakov (University of Birmingham, United Kingdom (Great Britain))
This paper presents a statistical analysis of the radar detection characteristics of FMCW based automotive radar in the presence of FMCW interference. The characteristics of interference are investigated and compared with Gaussian noise by determining the distribution and probability of detection for different waveform parameters. The majority of cases show that interference in the victim receiver has a low correlation value and a probability density function close to a Gaussian distribution, with some specific cases demonstrating a sinusoidal distribution. The analysis is validated by experimental results for specific case performance in a controlled laboratory environment.
Statistical Study of Hardware Impairments Effect on mmWave 77 GHz FMCW Automotive Radar
Mohammad Hossein Moghaddam and Sina Rezaei Aghdam (Chalmers University of Technology, Sweden); Alessio Filippi (NXP Semiconductors, The Netherlands); Thomas Eriksson (Chalmers University of Technology, Sweden)
In this paper, we analyze the effects of hardware impairments on 77GHz FMCW automotive radar performance. Joint in-phase/quadrature imbalance (IQI) and phase noise effects on frequency-modulated continuous-wave (FMCW) radar transceiver front-end are modeled through statistical analysis of distortion and noise. We derive the signal to distortion plus noise ratio, constant false alarm rate, and range-Doppler sensitivity analysis for both the joint and the individual effects of impairments and validate the formulations with simulations. The represented modeling and analysis can be used in millimeter wave (mmWave) FMCW automotive radar signal processing algorithms for optimum transceiver design.
Localization and 3D Mapping Using 1D Automotive Radar Sensor
Robin van Gaalen (TU delft, The Netherlands); Faruk Uysal (Delft University of Technology, The Netherlands); Alexander Yarovoy (TU Delft, The Netherlands)
This paper establishes novel methods for vehicle localization and mapping using a 1D linear automotive radar array in conjuncture with pre-existing lidar maps, and tests if the generated radar map can be made to be 3 dimensional. The basic design of this study was to implement a SLAM (Simultaneous Localization And Mapping) system that co-registers radar data to radar data, and/or register radar data to lidar data. After the execution of experiments, it was established that it possible to localize the car by relating observed radar data to pre-made lidar maps, and to continually add to a cumulative map made with the radar data that can further aid the localization process. Furthermore, the radar map created using the 1D linear automotive array can be extended to 3D with proposed processing chain, though more experiments to establish the full potential of this capability are recommended.
A 79GHz Millimeter Radar Based Car and Motorcycle Counter
Weijie Liu (Panasonic Corporation, Japan); Yoshiyuki Okubo (Panasonic, Japan); Jerry Sun and Yasuhiro Aoyama (Panasonic, Taiwan)
According to motorization evolution motorcycle ownership gets very high in many Asian cities so that a traffic counter applicable to motorcycles is necessary to improve road traffic environment. On the other hand a 79GHz millimeter radar is suitable for such an application as its high resolution and robustness at nighttime and in bad weather. In the paper we propose two methods to detect, classify and count cars (four-wheel vehicles) and motorcycles based on radar measurements of point cloud. One method is a target based approach where vehicles (cars and motorcycles) are extracted and tracked by grouping points and aggregated using a count line. Another method uses point cloud directly and aggregates with a count region. The methods are tested on a public road in Taiwan. The test results show that both methods perform with more than 90% accuracy when counting vehicles totally, but the second method is much better to count cars and motorcycles separately.
Automotive Radar Waveform Parameters Randomisation for Interference Level Reduction
Fatemeh Norouzian (University of Birmingham, United Kingdom (Great Britain)); Anum Ahmed Pirkani (University of Birmingham, United Kingdom (Great Britain) & AKSA-SDS, Pakistan); Edward Hoare and Mikhail Cherniakov (University of Birmingham, United Kingdom (Great Britain)); Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain))
With the rapid increase in the number of automotive radars on the road, interference becomes unavoidable. This paper analyses the effectiveness of dithering or randomization of FMCW modulation (applying random idle time jitter) on reducing interference. This analysis is undertaken for various interference scenarios and modelling-based results are compared. The results show that the efficiency of interference reduction by employing random idle time jitter strongly depend on the radar waveform parameters and the achievable reduction in interference level is in the range of 0 to 27 dB.
GIP Test for Automotive FMCW Interference Detection and Suppression
Thomas Pernstal (Flöjelbergsgatan 6 & SafeRadar Research, Sweden); Johan Degerman and Henric Broström (SafeRadar Research Sweden AB, Sweden); Vu Viet Thuy and Mats I. Pettersson (Blekinge Institute of Technology, Sweden)
This paper addresses the problem of mutual interference between automotive radars. The rapid growth of automotive and commercial radar systems on the market does not only facilitate new applications, e.g., advanced driver assistant systems, but also put demands on the possibilities for co-existence, i.e. cohabitant systems. For military radar systems, various jammer and interference mitigation methods have been extensively analyzed and evaluated for decades. However, until now, the co-existence and influence of jamming/interference have almost been ignored for the commercial radar business. A Generalized Inner Product, GIP, test based outlier detector and interference estimation is presented here, which suppress the interferences only in those Directions of Arrival, DOA, and time domain portions where the nuisance signals appear. We will denote this GIP test based Interference Detector and Suppression as the GIDS method. Using GIDS, the target detection performance for the specific interference DOA will merely have a small loss instead of being completely suppressed, e.g., sample matrix inversion implementation of spatial nulling. The proposed technique is robust and does not rely on any calibration for the interference cancellation. Based on simulation and experimental data, we have shown that without losing target detection performance, we achieved up to about 30 dB enhancement for the Signal to Interference and Noise Ratio.
Assessment of Micro-Doppler Based Road Targets Recognition Based on Co-Operative Multi-Sensor Automotive Radar Applications
Pasquale Striano (University of Strathclyde & Mr, United Kingdom (Great Britain)); Christos V. Ilioudis, Carmine Clemente and John J Soraghan (University of Strathclyde, United Kingdom (Great Britain))
Radar systems have become one of the principal sensory components in automotive vehicles, due to their ability to detect and discriminate between different objects in various scenarios. In this paper the micro-Doppler signature is used to identify road targets as cyclist, person, group of people, dog walking, and dog trotting. In order to boost the performance of Automatic Target Recognition in automotive environment, each node could share its micro-Doppler based features in a co-operative manner, using novel Vehicle To Vehicle communication frameworks based on joint radar and communication systems. The classification performance is evaluated considering two scenarios, a single-sensor scenarios where the micro-Doppler signature is observed by a single user, and a multi-sensor scenarios where each user shares its feature vector.
Interference Characterization in FMCW Radars
Sandeep Rao (Texas Instruments, India); Anil Varghese (Indian Institute of Science & Texas Instruments, India)
As the penetration levels of Automotive radars increase so will the problem of mutual interference. It is important to understand how this will affect the performance of current radar technology. The aim of this paper is the statistical characterization of interference in the context of fast chirp FMCW radar. We focus on two categories of interference: 'Parallel' and 'Sweeping'. Through analysis and simulation we provide an insight into the mechanism of interference and also the inbuilt resilience of FMCW radars which help in interference mitigation. Monte Carlo simulations (of multiple radars at randomized distances) are used to provide a sense of radar performance in real-life scenarios. We discuss the performance difference between real and complex base-band FMCW architectures. For comparison we also present simulation results for PMCW radars.

WA-P5: Experimental short-range target detection and classification

Chair: Pepijn B. Cox (TNO, The Netherlands)
Class Factorized Variational Auto-encoder for Radar HRRP Target Recognition
Jian Chen, Lan Du and Leiyao Liao (Xidian University, China)
In this paper, a class factorized variational auto-encoder (CFVAE) is developed for radar high-resolution range profile (HRRP) target recognition. The proposed model integrates the discriminative information into the deep statistical modeling of HRRP. In detail, the CFVAE utilizes an encoder to project all observations to the deep latent space and then separately defines the generative process from the latent features to observed data in each class, via the collection of specific class-decoders. Due to the stronger ability of each class-decoder to the description of observations in the corresponding class, the CFVAE model can directly assign a test sample into the class that corresponds to the decoder with the minimum reconstruction error, which avoids the mismatch of the extracted features and the back-end classifier. Meanwhile, compared to the traditional variational auto-encoder (VAE) describing the whole dataset with a single decoder, the proposed method has the capacity to give a more accurate description for all observations, thus beneficial to the improvement of features in characterization ability. Moreover, instead of imposing a fixed prior on the latent representations, our CFVAE model learns the conditional prior distribution based on the samples' labels, which further enhances the discrimination of the latent space. Finally, experiments on the measured HRRP dataset demonstrate the promising recognition performance of the proposed method.
From Unsupervised to Semi-Supervised Anomaly Detection Methods for HRRP Targets
Martin Bauw (MINES ParisTech, PSL Research University, CMM - Center of Mathematical Morphology & Thales LAS, France); Santiago Velasco-Forero (MINES ParisTech, PSL Research University, CMM-Center of Mathematical Morphology, France); Jesus Angulo (Mines ParisTech, France); Claude Adnet (Thales Air Systems, France); Olivier Airiau (Thales LAS, France)
Responding to the challenge of detecting unusual radar targets in a well identified environment, innovative anomaly and novelty detection methods keep emerging in the literature. This work aims at presenting a benchmark gathering common and recently introduced unsupervised anomaly detection (AD) methods, the results being generated using high-resolution range profiles. A semi-supervised AD (SAD) is considered to demonstrate the added value of having a few labeled anomalies to improve performances. Experiments were conducted with and without pollution of the training set with anomalous samples in order to be as close as possible to real operational contexts. The common AD methods composing our baseline will be One-Class Support Vector Machines (OC-SVM), Isolation Forest (IF), Local Outlier Factor (LOF) and a Convolutional Autoencoder (CAE). The more innovative AD methods put forward by this work are Deep Support Vector Data Description (Deep SVDD) and Random Projection Depth (RPD), belonging respectively to deep and shallow AD. The semi-supervised adaptation of Deep SVDD constitutes our SAD method. HRRP data was generated by a coastal surveillance radar, our results thus suggest that AD can contribute to enhance maritime and coastal situation awareness.
Micro-Range Micro-Doppler for Classification
Dave Tahmoush (University of Kansas, USA)
Micro-range micro-Doppler radar approaches can separate different parts of a radar signature in range as well as in Doppler. The differences between individual human signatures and the signatures of groups of humans walking together can then be recognized. Using micro-range micro-Doppler, the signature from a group of people walking together toward the radar can be distinguished from a single person walking toward the radar with approximately 89% accuracy.
Preliminary Studies to Identify Oil Well Perforations with Radar Installed Outside the Well
Yaser Norouzi, Ali Jamaluddin, Gholamreza Moradi and Sajjad Kaveh (Amirkabir University of Technology, Iran)
It is important to know if oil well perforations are open or closed in the well. With notice the closure of the oil well perforations, chemical or ultrasonic methods can be used to rinse the well to extend the life of the well. In recent years, the authors have suggested using radar to detect these perforations. To do this, they have to lower the radar to the bottom of the well and near the perforations. This works for wells with depths of several kilometers and temperatures of 350 ° C and pressure of 600MPa will be very difficult and expensive. In this article, we have investigated whether these perforations can be detected without sending radar into the well? Our calculations show that it is possible to identify holes with a diameter of 5 to 25 mm, up to a depth of 1,500 meters with modern technology, but more advanced methods are needed for further depths.
Comparative Analysis of 300 GHz and 79 GHz Radar Performance in Fire Environments
Aleksandr Bystrov, Liam Y. Daniel, Edward Hoare, Fatemeh Norouzian and Mikhail Cherniakov (University of Birmingham, United Kingdom (Great Britain)); Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain))
This paper presents an experimental comparative study of the propagation of microwave signals in the frequency ranges of 300 GHz and 79 GHz through fire. The operation of Low Terahertz imaging radar was investigated in various real scenarios, including fire with strong flame, dense smoke, and water vapor. The ability of Low Terahertz radar to ensure visibility of objects in fire environments was proven. In all scenarios, radar signal absorption was measured and in the case of steam was compared with theoretical calculations; the results are in good agreement. The analysis of the experimental results allows us to conclude that there are good prospects for Low Terahertz radar in the field of firefighting equipment for vision and navigation.
UAV Micro-Doppler Signature Analysis
Dave Tahmoush (University of Kansas, USA)
The radar phenomenology of UAVs interacts with the micro-Doppler signal processing in interesting and useful ways. This paper adjusts the micro-Doppler processing to establish more consistent UAV micro-Doppler signatures and improve the separability of relevant features. The outlined approach is developed in the context of UAV rotor analysis. The techniques are demonstrated in both simulated and experimentally measured results of UAV rotor blades to distinguish three different types of UAV.
Design and Development of K-Band FMCW Radar for Nano-Drone Detection
Safiah Zulkifli and Alessio Balleri (Cranfield University, United Kingdom (Great Britain))
Nano-drones, are insect-like size drones with a capability of intrusion to provide intelligence and potentially violate secure establishments and public privacy rights. Nano-drones are already an existing technology which is becoming more and more available, portable, affordable and easily operated. As such, they may soon become a plausible defence and security threat. This paper presents the design and development of a K-Band FMCW radar prototype for nano-drone detection. The FMCW radar prototype consists of connectorised components operating at a carrier frequency of 24 GHz and offer high parameter selection flexibility. Experiments have been carried out in order to evaluate the system performance. Results show that a small Arcade PICO Drone Nano Quadcopter (smaller than 5 cm) could be detected, and that its micro-Doppler signature could be extracted from data.
Simulation-Based Approach to Classification of Airborne Drones
Lasse Lehmann (Technical University of Denmark & Terma A/S, Denmark); Jørgen Dall (Technical University of Denmark, Denmark)
Recognition of drone type provides valuable information to assess the capability of drones, which is essential to airspace monitoring. Classification of drones on the basis of radar data is dominated by the use of supervised learning, which exploits different and often combined representations of the micro-Doppler signatures of the target. However, it is expensive and cumbersome building a catalogue of several drone micro-Doppler signatures using real data. We introduce a simulation-frame-work to generate radar data from point-scatterer targets, with associated radar cross section evaluated using physical optics. Small scale lab tests validate the fidelity of the simulated radar data, while the utility of the synthetic data for classification is tested using established methodology for classification.

WA-P6: Radar tracking and track-before-detect

Chairs: Florian Meyer (University of California San Diego, USA), Luca Pallotta (University of Roma Tre, Italy)
Ship Tracking in High-Resolution Range-Compressed Airborne Radar Data Acquired During Linear and Circular Flight Tracks
Sushil Joshi (Deutschen Zentrums für Luft- und Raumfahrt (DLR), Germany); Stefan V. Baumgartner (German Aerospace Center (DLR), Germany)
Ship tracking is a key maritime surveillance requirement. Research on ship tracking using airborne radar platforms is not widely available in the open literature. In this paper, a concept of ship tracking using range-compressed (RC) airborne radar data is proposed. The proposed tracking algorithm is suitable for dense multi-target scenarios. Tracking is performed in the range-Doppler domain where the moving ships may appear out of the clutter region, improving their detectability. A powerful track management system is developed using SQLite database for handling gaps in the target detections and also the false detections. Some simulations and real experimental results from DLR's F-SAR system are provided to prove the concept.
A Data Association Approach for Plot-Sequences Outputted by Multi-Frame Track-Before-Detect
Kezhu Liu, Ph Zhang, Wujun Li and Wei Yi (University of Electronic Science and Technology of China, China); Lingjiang Kong (University of Electronic Science and Technology of China (UESTC), China)
This paper addresses the data association problem with respect to the multi-frame track-before-detect (MF-TBD). Different from the classical target detection and tracking approaches which declare target and update tracks for each single frame, the MF-TBD can achieve superior performance by integrating target energy over several consecutive frames, especially for dim or fluctuating targets. Moreover, with jointly processing consecutive frames of raw data, the outputs of MF-TBD are plot-sequences. However, in the single-target or multi-target scenario, there needs to assign the plot-sequences to the existing trajectories. In other words, MF-TBD is not a complete target detection and tracking system, or rather, the further tracking treatments are needed, such as the association step. In this paper, by referring to the existing works of data association methods, a novel data association approach for the plot-sequences outputted by MF-TBD is developed. Last, the simulation results show that the proposed algorithm can correctly associate target trajectories with the corresponding plot-sequences and significantly enhance the tracking performance compared with the traditional tracking algorithm.
Trajectory Tracking by the Interacting Multiple Model Algorithm: Genetic Approach to Improve the Performance
Dmitrii A Bedin and Alexey Ivanov (N. N. Krasovskii Institute of Mathematics and Mechanics, Russia)
In the paper, we consider an application of the genetic approach to improve the performance of the trajectory tracking procedure. Our aim is to reduce the horizontal position estimation errors. The procedure is based on the well known Interacting Multiple Model algorithm and has many parameters to be adjusted, e.g., elements of the transition probability matrix. A genetic tuning algorithm is elaborated and some numerical experiments on simulated aircraft trajectories are made.
Unbiased Measurements Conversion Based Sequential Filtering for Target Tracking with Range, Range Rate and Direction Cosine Measurements
Lifu Li (University of Electronic Science and Technology of China, China); Ting Cheng (University of Electronic Science and Technology, China)
The measurements in phased array radar are obtained in the direction cosine coordinates, which include the range and direction cosine measurements. In order to realize the target tracking with nonlinear measurements and improve the tracking performance with range rate information, the unbiased measurements conversion based sequential filtering for target tracking with range, range rate and direction cosine measurements is proposed. First, the position measurements are processed by the converted measurements Kalman Filter (CMKF), where the statistical characteristics of the converted measurements error based on range and direction cosine measurements are calculated based on the predicted information to remove the correlation between the covariance matrix and the measurement noise; Based on the filtering result from position information filter, the pseudo measurement, which is constructed by the range and range rate measurements, is input to realize sequential filtering to improve the position filter result. The statistical characteristics of pseudo measurement are also calculated based on the predicted information. Simulation results demonstrate the effectiveness of the proposed algorithm.
Joint Detection and Tracking Scheme for Target Tracking in Moving Platform
Yongsheng Guan (Information Science Academy of China Electronics Technolog Group Corporation, USA); Yingping Wang (Xidian University, China)
In this paper, a joint detection and tracking (JDT) scheme is presented for single target tracking in clutter using a moving radar platform. The core of the JDT scheme is to adaptively modify the detection threshold according to the feedback information from the tracker to the detector, while guaranteeing the averaged false alarm rate in the validation gate is a constant. The feedback information we utilized is the target's predicted location distribution, which can be achieved through jointly estimating platform and target states. Simulation results suggest that the JDT scheme can provide improved detection and tracking performance significantly, when compared with the standard probabilistic data association method.
Tracker Integrity Performance Assessment
Nicolas Honoré and Marc Fragu (Thales LAS/AMS, France); Christophe Labreuche and Bruno Marcon (Thales Research and Technology, France)
In Air Traffic Management, tracking components are key to optimize traffic, by ensuring spatial separation between airplanes. The aim of the work is to propose the monitoring of the tracking system though a tracker integrity performance assessment. It is based on two components. The first one computes in quasi-real time tracking Key Performance Indicators. The second one aggregates the mandatory and recommended requirements from the ESASSP norm, and the preferences of the Air Controllers on the priority among these requirements, to construct a single score representing the overall tracking quality of service. We show that these two types of requirements must be handled separately.
Track-Before-Detect Method Based on Preprocessing with Pseudo-Spectrum
Liangliang Wang and Gongjian Zhou (Harbin Institute of Technology, China); Thia Kirubarajan (McMaster University, Canada)
Traditional dynamic programming based track-before-detect (DP-TBD) methods may suffer performance loss due to the suboptimal preprocessing. In this paper, a DP-TBD method based on preprocessing with pseudo-spectrum is proposed. For each cell in the observed image, a pseudo-spectrum is constructed around the cell itself with measurement value as its peak according to a point spread function. Samples of the pseudo-spectrum are added onto corresponding cells in the current frame. This facilitates improved performance through the intraframe integration by using spilled target echo energy. Then, dynamic programming procedure is performed for multiframe integration. The pseudo-spectrum based preprocessing is presented, and energy integration procedure for the proposed DP-TBD is provided in detail. Theoretical analysis and simulation result demonstrate the validity of the proposed method.
Supervised Learning Based Online Filters for Targets Tracking Using Radar Measurements
Jie Deng, Wei Yi, Kai Zeng, Qiyun Peng and Xiaobo Yang (University of Electronic Science and Technology of China, China)
In the field of the radar target tracking, the state filtering plays an important role in estimating the target state. One of the widely-adopted filter is the Bayesian filter, which requires the prior information and an accurate modeling of the real tracking scene. Thus, the matching degree of the dynamic model has a key impact on the state estimation accuracy of the Bayesian filter. However, the target dynamic and radar measurement models cannot be approximated perfectly in an unknown and complicated environment, and the state estimation accuracy of the model-based filter is limited. To address the limitations of the Bayesian filter, a supervised learning based online filter for target tracking is proposed in this paper. In the proposed filter, a mapping among the radar measurements is first established in the context of the polar coordinate system. Then, based on data-driven, the state filtering is directly implemented to obtain the state estimate by using this mapping relationship. As such, the prior information is not required in the proposed filter, hence the proposed filter can have a good estimation accuracy in unknown and complicated environments. Finally, simulation experiments clarify the effectiveness of our proposed filter via comparing with the traditional filter.
A Multi-Scan Joint Tracking and Classification Method for Weak Targets
Kai Zeng, Wei Yi, Qiyun Peng and Jie Deng (University of Electronic Science and Technology of China, China)
This paper proposes a novel multi-scan algorithm with the task of the joint tracking and classification (JTC) for weak targets to address the limitations existed in the traditional single-scan JTC algorithms based on the threshold detection. At each scan, the measurements both from the radar system and the electronic support measures (ESM) sensor, which is one of the passive and bearing-only sensors, are first adopted. Then, by using these measurements of multiple scans, a fast multi-scan JTC (MS-JTC) algorithm is proposed. It is shown that the MS-JTC can bring about a better signal to noise ratio (SNR) for radar systems and better classification due to the accumulation information of multiple scans. Finally, by comparing with the existing single-scan-based algorithm, simulation experiments are executed to validate the effectiveness of our proposed algorithm.
An Aircraft-Centered Multi-Frame TBD Method for Airborne Radar Systems
Qiyun Peng, Wei Yi, Kai Zeng, Jie Deng, Wujun Li and Lingjiang Kong (University of Electronic Science and Technology of China, China)
It is widely verified that the multi-frame track-before-detect (MF-TBD) is effective for the detection and tracking of weak targets. This paper considers the extension of MF-TBD to moving surveillance platform, specifically, the airborne radar system. For static radars, MF-TBD needs to first map the received raw data to an absolute coordinate, mostly the WGS-84 coordinate, and then perform the raw data integration between multi-frame to accumulate target energy along physically feasible states. The time-varying surveillance space and biased complex coordinate system conversion, caused by the aircraft motion, produce the unaligned multi-frame mapping. It brings the cell misalignment and dislocation error that contribute to the challenge of the tracking accuracy in the absolute coordinate. To address these limitations, the proposed aircraft-centered multi-frame TBD (AC-MF-TBD) method adopts the concept of relative kinematics making the aircraft as the center of each measurement frame and doing raw data integration in this relative coordinate, namely the inertial coordinate. It makes the mapping no longer time-varying and develops the tracking accuracy on the airborne radar. AC-MF-TBD essentially avoids the errors introduced by the aircraft motion in the absolute coordinate. In addition, the coordinate conversion using the attitude information of the aircraft is still a necessary process for AC-MF-TBD. The attitude angle calibration is used to improve the conversion accuracy. The simulation shows the performance of the proposed algorithm.
Target Tracking and Ghost Mitigation Based on Multi-view Through-the-wall Radar Imaging
Huquan Li (University of Electronic Science and Technology of China, China); Guolong Cui (University of Electronic Science and Technology of China (UESTC), China); Shisheng Guo (University of Electronic Science and Technology of China, China); Lingjiang Kong (University of Electronic Science and Technology of China (UESTC), China); Xiaobo Yang (University of Electronic Science and Technology of China, China)
Multiple human target tracking in an enclosed structure is a critical mission for through-the-wall radar imaging (TWRI). Current efforts monitor the surveillance area from a particular view, where the ghost targets caused by the multipath and shadow effects would introduce numerous fake tracks. In this paper, we deal with the problem of target tracking exploiting multi-view TWRI. A sequential filtering framework is utilized to estimate the tracks of both the human and ghost targets based on the measurements from multiple radar nodes. Subsequently, a ghost mitigation method is proposed based on the view-dependent features of the ghosts. Finally, the proposed algorithm is validated by numerical simulations.

Thursday, September 24

Thursday, September 24 9:00 - 10:40 (Europe/Rome)

ThM-L1: DOA Estimation

Room: Ch1
Chairs: Vincenzo Carotenuto (University of Naples Federico II, Italy), Lingjiang Kong (University of Electronic Science and Technology of China (UESTC), China)
9:00 A Comparison of Two Stochastic Maximum Likelihood DoA Estimators - Rotating Array Case
Michał Meller (Gdansk University of Technology & PIT-RADWAR S.A., Poland); Kamil Stawiarski (Gdansk University of Technology, Faculty of Electronics, Telecommunications and Computer Science & PIT-RADWAR S.A., Poland)
The accuracy of two variants of stochastic maximum likelihood direction of arrival estimators is compared under the assumptions of a rotating array and two sources. The difference between the estimators lies in how they approach the source correlation. One estimator treats sources as uncorrelated, while the other includes the source correlation coefficient as a parameter to be estimated. Analytic formulas for the mean squared error (MSE), worst-case MSE, and average MSE of both estimators are derived and verified using computer simulations. The simpler estimator is shown to be competitive to the more complex one for small and moderate levels of source correlation.
9:20 Robust Semiparametric DOA Estimation in Non-Gaussian Environment
Stefano Fortunati (CentraleSupélec, France); Alexandre Renaux (Paris-Sud University, France); Frederic Pascal (CentraleSupélec, France)
A general non-Gaussian semiparametric model is adopted to characterize the measurement vectors, or snapshots, collected by a linear array. Moreover, the recently derived robust semiparametric efficient R-estimator of the data covariance matrix is exploited to implement an original version of the MUSIC estimator. The efficiency of the resulting R-MUSIC algorithm is investigated by comparing its Mean Squared Error (MSE) in the estimation of the source spatial frequencies with the relevant Semiparametric Stochastic Cramér-Rao Bound (SSCRB).
9:40 Simultaneous Radar Detection and Constrained Target Angle Estimation via Dinkelbach Algorithm
Massimo Rosamilia (University of Naples Federico II, Italy); Augusto Aubry (Universita degli studi di Napoli, Italy); Antonio De Maio (University of Naples "Federico II", Italy); Stefano Marano (University of Salerno, Italy)
Simultaneous target detection and angle estimation with a multichannel phased array radar system is addressed in this paper. Starting from a linearized expression for the array steering vector around the beam pointing direction, the problem is cast as a composite binary hypothesis test where the unknowns, under the alternative hypothesis, include the target direction cosines displacements with respect to the array nominal steering. The problem is handled via the Generalized Likelihood Ratio (GLR) criterion where a decision statistic leveraging the Maximum Likelihood Estimates (MLEs) of the parameters is compared to a detection threshold. If crossed, target presence is declared and MLEs of the aforementioned displacements directly provide target angular position with respect to the pointing direction. From the analytic point of view ML estimation requires the solution of a constrained fractional quadratic optimization problem whose optimal solution can be found via Dinkelbach's algorithm. The analysis of the proposed architecture is developed in terms of detection performance and angular estimation accuracy also in comparison with some counterparts available in open literature and benchmark limits.
10:00 Large Array DOA Estimation Based on Extreme Learning Machine and Random Matrix Theory
Anqi Zhao, Hong Jiang and Qi Zhang (Jilin University, China)
Estimation of the direction-of-arrival (DOA) in large array systems owns wide prospect in radar applications. However, the traditional subspace DOA estimation algorithm has deteriorated performance when the numbers of antenna elements and samples grow in the same rate. Also, they are poorly adapted to the actual low signal-to-noise ratio (SNR) environment. In this paper, we investigate the large array DOA estimation based on extreme learning machine (ELM) and random matrix theory (RMT). Conventionally, ELM requires the activation function to be infinitely differentiable, which may lead to slow training rate for large arrays. According to RMT, it is proved that ELM with ReLU function as its activation function has asymptotic convergence in large dimension regime. Thus, a ReLU-ELM method for large array DOA estimation is put forward. Numerical simulations show that under low SNR, it has better performance than the traditional algorithms. Compared with the least square support vector machine (LSSVM) method, the ReLU-ELM algorithm can greatly reduce the training and testing time and improve the learning efficiency under the premise of ensuring excellent estimation performance.
10:20 Multi-frame DOA Estimation Algorithm for Weak Moving Targets
Ph Zhang, Kezhu Liu, Wujun Li, Wei Yi and Xiaobo Yang (University of Electronic Science and Technology of China, China)
The positioning ability of the DOA estimation algorithms in array signal processing depends on the correct selection of snapshot data including the target echo signal. However, in the scene with low signal-to-noise ratio (SNR) , the CFAR method, which is frequently used before DOA estimation, cannot effectively discover the snapshot data containing the targets, resulting in the loss of the target. In this paper, we propose a multi-frame DOA (MF-DOA) estimation algorithm, which implements multi-frame energy accumulation and snapshot data extraction based on the introduced target motion model so that the weak moving targets can be effectively located. The simulation data and real MIMO radar data are processed, and the results show that the MF-DOA algorithm has good ability to excavate weak moving targets and extend the radar detection range.

ThM-L2: Joint radar and communications

Room: Ch2
Chairs: Yonina C. Eldar (Weizmann Institute of Science, Israel), Marco Lops (University of Naples Federico II & CNIT - Consorzio Universitario Nazionale per le Telecomunicazioni, Italy)
9:00 Realization of a Joint MIMO Radar and Communication System Using a PSK-LFM Waveform
Muge Bekar, Christopher J. Baker and Edward Hoare (University of Birmingham, United Kingdom (Great Britain)); Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain))
In this paper, a new technique which enables communication data to be embedded into MIMO radar signal is presented. A linear frequency modulated (LFM) signal is used for radar purposes, and multiple phase shift keying (PSK) symbols, for communications, are embedded into the LFM signal. In this way, multiple communication symbols per radar pulse are generated. Such waveforms are subsequently combined within a multiple-input multiple-output (MIMO) configuration. Orthogonality between transmitters is ensured using time-division multiplexing (TDM). The performance of this novel PSK-LFM technique is demonstrated through both simulation and experimentation.
9:20 Effect Analysis of Spatial Modulation on Clutter Mitigation for Joint RadCom Systems and Solutions
Xin Zhang and Xiangrong Wang (Beihang University, China); Elias Aboutanios (University of New South Wales, Australia)
The increasing competition over the scarce spectrum has intensified the problem of spectrum congestion. Joint radar-communication (RadCom) systems, where radar and communication systems are integrated into one platform, have recently commanded significant attention for their ability to efficiently utilize the limited spectral resources. Major challenges of joint RadCom systems include shared waveform design, leading to the proposal of different signalling strategies. Among these, spatial modulation embeds random communication information into the complex beampattern of the shared waveform. The variation of transmit beampattern in a coherent processing interval adversely affects the Doppler coherency of clutter, in turn impacting the radar sensitivity. In this work, we analyse quantitatively the effect of beampattern variation on clutter modulation for joint RadCom systems. Then, we propose a calibrated matched filter based on subspace projection by fully utilizing the properties of spatial modulation. Simulation results demonstrate that the calibrated matched filter is effective at mitigating ground clutter without affecting the dual functions of joint RadCom systems.
9:40 Green Communications with Radar Spectrum Sharing
Emanuele Grossi (University of Cassino and Southern Lazio & Consorzio Nazionale Inter-universitario per le Telecomunicazioni (CNIT), Italy); Marco Lops (University of Naples Federico II & CNIT - Consorzio Universitario Nazionale per le Telecomunicazioni, Italy); Luca Venturino (Universita' degli Studi di Cassino e del Lazio Merdionale & Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), Italy)
In this paper, we consider a multiple-input multiple-output communication system and a surveillance radar sharing the same bandwidth. In this coexistence framework, we undertake a joint system design, where, different from previous studies, we adopt the energy efficiency (i.e., the amount of information reliably delivered per unit of consumed energy) as a figure of merit to be maximized at the communication system side. The radar performance is instead safeguarded by imposing a constraint on the minimum signal-to-disturbance ratio for each inspected range-azimuth resolution cell. The degrees of freedom for system optimization are the transmit powers of both systems, the space-time linear communication codebook, and the radar receive filters are the degrees of freedom for joint system optimization. The block coordinate ascent method is used to find an approximate solution to this problem, and a numerical example is provided to show the merits of the proposed design strategy.
10:00 Distributed Radar-Aided Vehicle-to-Vehicle Communication
Canan Aydogdu (Chalmers University of Technology, Sweden); Fan Liu and Christos Masouros (University College London, United Kingdom (Great Britain)); Henk Wymeersch (Chalmers University of Technology, Sweden); Mats Rydström (Veoneer, Sweden)
Establishing high-rate vehicle-to-vehicle (V2V) links with narrow beamwidth is challenging due to the varying network topology. A too narrow beam may miss the intended receiver, while a too broad beam leads to SNR loss. We propose to harness the high accuracy of radar detections to establish V2V links. In particular, we develop a distributed method where each vehicle associates local radar detections with GPS information communicated by nearby vehicles. The method relies on the transformation of relative to global coordinates, the definition of a suitable metric, and solving an optimal assignment problem.We demonstrate that the proposed approach avoids time-consuming channel estimation and provides high SNR, under the condition that reliable relative and absolute location information is present.
10:20 Can Automotive Radars Form Vehicular Networks?
Canan Aydogdu and Henk Wymeersch (Chalmers University of Technology, Sweden); Mats Rydström (Veoneer, Sweden)
Radar communications (RadCom) is a spectrally efficient way for removing automotive radar interference and thereby enhancing reliable radar sensing, via a single hardware for both radar and communications. When interference coordination does not use all the RadCom resources, opportunities for communicating additional data arise. We propose a new communication protocol, termed RadNet (for radar network), which forms a vehicular ad-hoc multi-hop network by automotive radars in a distributed manner. Simulation results obtained for high-way use cases show that RadNet can enable several Mbps data links without degrading the radar performance.

ThM-L3: Radar technology

Room: Ch3
Chairs: Antonio De Maio (University of Naples "Federico II", Italy), Jianyu Yang (School of Electronic Engineering, China)
9:00 Airborne Polarimetric Doppler Phased Array Weather Radar: Performance Requirements and Design Specifications
Jothiram Vivekanandan, Luis Perez-Clifford, Eric Loew and Adam Karboski (National Center for Atmospheric Research, USA)
Performance requirements for an airborne weather radar are more stringent than ground-based radar. Since the airborne radar has only a limited time for collecting measurements over a specified region, phased array radar with active electronic scanning array (AESA) enables rapid scanning of the antenna beam for collecting high-temporal and spatial resolution, research quality Doppler and polarimetric radar measurements. In the case of an AESA when a beam is steered electronically away from the broadside, the gain and beamwidth change as a function of scan angle and cross-coupling between dual-polarization sources occur. For a specified aperture size and solid-state amplifier, pulse compression is used to enhance the sensitivity of the received signal. To achieve reasonable along-track spatial resolution, the use of beam multiplexing (BMX) is essential. Beam multiplexing reduces errors in radar measurements while providing rapid updates of scan volumes. A beamforming architecture that supports BMX, and high-speed serial interconnect for frontend signal processing is essential for optimal performance of PAR. This paper describes the engineering design specifications of an antenna, transmit waveform, and weather radar signal processing subsystems.
9:20 Quantum Radar with Digital Waveform Generators
Fred E Daum (Raytheon, USA)
Quantum radars at X-Band suffer from two major issues today: cost and performance. The cost of quantum radar at X-Band is roughly ten orders of magnitude more than the corresponding classical radar today. Also, the detection performance of experimental quantum radars at X-Band is not any better than the optimal classical radar today. We discuss future research in X-Band quantum radars, including a theoretical idea to simulate quantum radar without cryogenic dilution refrigerators. This idea uses digital waveform generation rather than analog signal generation.
9:40 Multi-Resolution and Multi-Rate UWB ESM Receiver Design via Direct RF Sub-Nyquist Sampling
Abdullah Pinarcik (ASELSAN A. S., Turkey); Ali Bugra Korucu and Yasar Kemal Alp (ASELSAN Inc., Turkey); Lutfiye Durak-Ata (Istanbul Technical University, Turkey)
Design and implementation of ultra-wideband receivers are critical in Electronic Warfare (EW) systems. In this work, we propose a new multi-resolution, multi-rate, directRF sub-Nyquist sampling ultra-wideband Electronic Support Measures (ESM) receiver architecture. Multiple ADCs operating at different sampling rates are used for unambiguously resolving 9 GHz bandwidth. Multiple time and frequency resolutions are achieved by computing multiple FFTs with different lengths. The receiver is based on direct RF signal sampling, hence it has a simple hardware implementation. The proposed receiver has 9 GHz instantaneous bandwidth, and its instantaneous dynamic range is over 50 dB along this band. The total data rate is only 4.77 GHz, which is around one quarter of the theoretical 18 GHz Nyquist limit. Collected measurements from the implemented receiver hardware are presented in detail.
10:00 Sectoral Horn and Patch Array Antenna Design for a 76-81 GHz High-Resolution Tomographic Radar
Wafa Pathan Rahamtulla (Chalmers University of Technology); Andreas Och (Infineon Technologies Austria AG, Austria & Friedrich-Alexander University of Erlangen-Nuremberg, Germany); Patrick Hölzl (Infineon Technologies Austria AG); Stefan Schuster (voestalpine Stahl GmbH); Venkata Pathuri Bhuvana (Silicon Austria Labs GmbH, Austria); Robert Weigel (Friedrich-Alexander Universität Erlangen-Nürnberg, Germany)
Nondestructive testing is employed in a wide range of industrial applications for quality control and process optimization. Recently, tomographic millimeter-wave radar has been investigated for highly demanding scenarios such as imaging of gas distributions. In such systems, application specific antenna design is crucial to ensure a sufficient number of signal paths between the sensors while at the same time reducing unwanted reflections. This paper describes the design of both a sectoral horn and a patch array antenna producing a fan-shaped beam in the frequency range of 76-81 GHz. Prototypes are realized utilizing innovative additive manufacturing techniques and characterized in a semi-anechoic chamber. The impact of mounting the antennas behind quartz glass is investigated in both simulation and measurement. Finally, the antenna performance is compared with commercial standard gain horn antennas in a tomographic radar setup demonstrating a significant improvement in the quality of the reconstructed permittivity distribution.
10:20 Design of a Substrate Integrated Waveguide Slots Antenna in W Band for Aircraft Radar Application
Lamia Sadaoui (Universite Cote d'Azur, CNRS, LEAT, France); Jerome Lanteri (Université Nice Sophia Antipolis, France); Jean-Yves Dauvignac (Université Côte d'Azur, CNRS, LEAT, France); Claire Migliaccio (Université Côte d'Azur, CNRS, France); Fabien Ferrero (University Nice Sophia Antipolis, CNRS, LEAT & CREMANT, France); Laurent Brochier (Université de Nice-Sophia Antipolis, France); Ronnie Rexhaeuser and Thilo Lenhard (InnoSenT GmbH, Germany); Federico Lolli and Claudio Palleschi (Interconsulting, Italy)
The paper presents a design of a SIW (Substrate Integrated Waveguide) slot's antenna in W band for a sense-and-avoid radar module to tackle an emerging need of security during landing and take-off for small aircrafts. This work is developed inside the ODESSA project, funded by the European Union. The antenna was formed by 7 slots etched on the top face of a SIW structure designed on a Rogers RO3003G2 substrate. Simulation and experimental results, for both cases with and without radome, are presented inside 76-80 GHz frequency range. Good performances are obtained for efficiency and realized gain which were the main objectives of this work.

ThM-L4: Radar phenomenology and modeling

Room: Ch4
Chairs: Luke Rosenberg (Defence, Science and Technology Group & University of Adelaide, Australia), Simon Watts (UCL, United Kingdom (Great Britain))
9:00 Scanning Radar Simulation in the Maritime Environment
Luke Rosenberg (Defence, Science and Technology Group & University of Adelaide, Australia); Simon Watts (UCL, United Kingdom (Great Britain)); Stephen Bocquet (DST Group, Australia)
Modelling and simulation of radar sea clutter is extremely important to assist with radar system design, evaluate radar detection algorithms and stimulate radar processors during development and testing. In this paper, the key tradeoffs in designing a scanning maritime radar are presented with details of how the radar return can be simulated. The sea clutter model is a compound Gaussian model with the spectrum modelled using the evolving Doppler spectrum model with an appropriate representation of the texture. This is important as the Doppler spectrum plays a key role in determining the statistics that are important for coherent processing. All parameters used in the simulation are determined by empirical models which in turn rely on user defined inputs including sea state, the wind and swell direction and the grazing angle.
9:20 HF RCS of Small Boat Displacement in the Ocean
Gordon Frazer and Charlie Williams (FrazerLab Pty Ltd, Australia)
We present a numerical modelling approach for determining the HF radar cross-section of a volume displacement in the ocean surface. This displacement, for example, might be from a non-conductive wooden-hulled boat with no metal structure or fittings. Such a vessel typically has too small a radar cross-section to be detected considering only the scattering properties of the boat in isolation. Our modelling technique is approximate but indicates wooden, or other non-conductive material boats may be detectable by skywave radar in some instances, solely due to their displacement in the ocean. Our results suggest new radar management strategies to improve the likelihood of detecting these targets. However, we continue to require the preferred ship detection criteria of low sea-state and low doppler-spread ionospheric propagation conditions. We also assume the boat is moving at speed and heading combinations such that the target Doppler is separated from the ocean first- order Bragg scatter.
9:40 Multistatic Radar Measurements of UAVs at X-Band and L-Band
Piers Jerome Beasley, Matthew Ritchie, Hugh Griffiths and William Miceli (University College London, United Kingdom (Great Britain)); Michael R Inggs (University Cape Town, South Africa)
This paper presents analysis of data captured with the NeXtRAD multistatic radar system during a fortnight of experimental trials in December 2019. The trials saw, for the first time, the NeXtRAD system capturing interleaved X-band and L-band measurements of multiple UAVs in simultaneous monostatic and bistatic configurations. Analysis is presented of the UAV's micro-Doppler signatures, permitting a discussion into the challenges some UAV platforms present for reliable detection. Comparisons are also made between X-band and L-band monostatic and bistatic UAV radar backscatter allowing conclusions to be drawn over the benefits of particular radar configurations for aiding UAV detection.
10:00 Bi-static Reflectivity Patterns of Vulnerable Road Users in the C-V2X Frequency Range
Andreas Schwind, Willi Hofmann, Sreehari Buddappagari Jayapal Gowdu and Ralf Stephan (Technische Universität Ilmenau, Germany); Reiner S. Thomä and Matthias Hein (Ilmenau University of Technology, Germany)
Automated and connected driving is a key technology for the future mobility. The sensor data fusion of different cooperating mobile nodes will become an essential factor to achieve this goal. Current developments show an enormous potential to increase the environmental awareness through joint communication and radar sensing. In this respect, future radio channel models aim to include bi-static reflectivities and radar cross-sections of road users in the nearfield as well as the far-field. This paper presents indoor measurements of the bi-static reflectivity of vulnerable road users between 1 GHz and 10 GHz, using a bicycle as an example. The measurement setup using two independent turntables covers the entire angular range of illumination and observation angles. Suitable signal post-processing methods, i. e., phase-coherent background subtraction and time-domain gating, minimize the antenna crosstalk and unwanted reflections. The results are compared with electromagnetic full-wave simulations and are evaluated in time and frequency domain. The resulting data show that, depending on the illumination and observation angles, the bi-static reflectivity of the bicycle is maximum around the forward-scattering region and around specular reflection over the entire frequency range.
10:20 Virtual Multistatic Illumination by Exploitation of Multipath Propagation with Coherent MIMO Radar
Oliver Biallawons (Fraunhofer FHR, Germany); Joachim Ender (University of Siegen, Germany)
In this paper the first results with the coherent MIMO radar MIRA-CLE Ka are presented in which multipath propagation is exploited. With the separation of the direct- and indirect paths from and to the radar a virtual multistatic illumination can be obtained. Due to this method the amount of information can be increased and the following processing or classification can be improved or robustified.

ThM-SS13: SAR meets AI

Room: Ch5
Chairs: Mario Costantini (E-GEOS - an Italian Space Agency and Telespazio Company, Italy), Giuseppe Scarpa (Università "Federico II" di Napoli, Italy)
9:00 Balanced Feature Pyramid Network for Ship Detection in Synthetic Aperture Radar Images
Tianwen Zhang, Xiaoling Zhang, Jun Shi, Shunjun Wei and Jianguo Wang (University of Electronic Science and Technology of China, China); Jianwei Li (Naval Aeronautical University, China)
Ship detection in Synthetic Aperture Radar (SAR) images is a fundamental but challenging task. Nowadays, given that the huge imbalance between sparse-distribution ships and complex backgrounds in training process, most existing deep-learning-based SAR ship detectors often face great difficulty in further improving accuracy. Therefore, to solve this problem, in this paper, a novel Balanced Feature Pyramid Network (B-FPN) is applied to enhance detection accuracy. Different from the raw Feature Pyramid Network (FPN), B-FPN utilizes the same-deep integration balanced semantic features to strengthen the multi-level features in the feature pyramid, by means of four steps, namely rescaling, integrating, refining and strengthening, which do not increase too much network parameter quantity. Experimental results on the open SAR Ship Detection Dataset (SSDD) shows that B-FPN can make a 7.15% mean Average Precision (mAP) improvement than FPN.
9:20 InSAR Phase Unwrapping Using Convolutional Neural Network
Francesco Calvanese (University of Naples Federico II, Italy); Francescopaolo Sica (German Aerospace Center (DLR), Germany); Giuseppe Scarpa (Università "Federico II" di Napoli, Italy); Paola Rizzoli (German Aerospace Center (DLR), Germany)
Bi-dimensional phase unwrapping is among the main critical tasks in SAR interferometry. Indeed, before the actual topography or deformation retrieval, the absolute phase values should be reconstructed from their modulo-2π wrapped version. Due to the presence of noise, the interferometric phase normally presents residues, i.e. phase jumps greater than π on a single pixel. The residues imply that the unwrapping procedure is path-dependent, i.e. it admits different solutions. In this work, we present a preliminary investigation for the implementation of a phase unwrapping algorithm that exploits both the interferometric phase and coherence as input to a Convolutional Neural Network. The obtained results are compared with state-of-the-art algorithms.
9:40 An Adversarial Learning Approach for Oil Spill Detection from SAR Images
Federico Ronci, Corrado Avolio, Mauro Di Donna and Massimo Zavagli (E-Geos, Italy); Veronica Piccialli (University of Rome Tor Vergata, Italy); Mario Costantini (E-GEOS - an Italian Space Agency and Telespazio Company, Italy)
Oil spills, caused by accidents or by ships cleaning their tanks, represent big threats for maritime and coastal ecosystems health. A very effective detection of oil spills can be performed using satellite synthetic aperture radar (SAR) systems, operating regardless of cloud coverage and sunlight and capable of discriminating oil from regular sea surface. However, discriminating between real oil spills and look-alikes (such as natural oils and seepages, often occurring in upwelling sea areas), although well performed by expert SAR image interpreters, poses a great challenge for automatic processes. In addition, a visual check performed by human operators on a great number of images would be too expensive. Therefore, many solutions for automatic detection have been tried in the last few years, using probabilistic models and, more recently, machine learning. This work presents an innovative solution based on image-to-image translation using convolutional neural networks (CNNs) trained with an adversarial loss function. The proposed approach has been tested, with very promising results, using Radarsat-2 and Sentinel-1 SAR data over the Mediterranean Sea and some areas of the Atlantic Ocean and the North Sea.
10:00 Low Resolution for DNN in SAR
Iulia Calota and Daniela Faur (University Politehnica of Bucharest, Romania); Mihai Datcu (German Aerospace Center, Germany)
In this paper, we study the impact of resolution on the classification of SAR images. The motivation for our work is comprised in just one question: Do we need high-resolution images for a good classification or can we use a sensor with lower resolution? High-resolution sensors, for either multispectral image or SAR image retrieval are expensive. Moreover, the high-resolution of the image is translated in large loads of data to be stored. In our study, we train a convolutional neural network with SAR images of different spatial resolutions and compare the classification results, in order to specify a range of classes that do not need high resolution to be recognized. We then use our own dataset with high-resolution SAR images that contain similar classes to perform a second classification. We show that a resolution 4 times smaller than the initial one can still achieve good results.
10:20 Classification of Synthetic Aperture Radar Images of Icebergs and Ships Using Random Forests Outperforms Convolutional Neural Networks
William Franz Lamberti (George Mason University, USA)
Synthetic aperture radar (SAR) is a common technique for capturing vessels and icebergs on the ocean surface. Convolutional Neural Networks (CNNs) are a popular approach to interpret classes captured in images which include ships and icebergs. However, CNNs are difficult to explain and are computationally expensive. In this paper, we built a random forest (RF) model which outperforms CNN based approaches by 7% and 11% on the testing and validation data, respectively. The RF model used interpretable metrics. These powerful metrics provide insight to what is important to distinguish the two classes from one another. Thus, despite noise present in the SAR images, the RF model was able to provide meaningful classifications between ships and icebergs.

Thursday, September 24 10:40 - 11:00 (Europe/Rome)

Coffee Break

Thursday, September 24 11:00 - 12:40 (Europe/Rome)

ThM-L10: High resolution radar imaging

Room: Ch5
Chairs: Michail Antoniou (University of Birmingham, United Kingdom (Great Britain)), Debora Pastina (Uniroma, Italy)
11:00 First Results of a Joint Measurement Campaign with PAMIR-Ka and MIRANDA-94
Ingo Walterscheid, Patrick Berens, Michael Caris, Stefan Sieger, Olaf Saalmann, Daniel Janssen, Gabriel El-Arnauti and Angel Ribalta (Fraunhofer FHR, Germany); Daniel Henke (University of Zurich, Switzerland); Elias Mendez Dominguez (University of Zurich & Remote Sensing Laboratories, Switzerland)
Fraunhofer FHR has participated in the international measurement campaign of the NATO research task group SET-250 with two airborne SAR systems in July 2019. The general objective of the trials was to investigate the use of multidimensional radar to increase the performance of radar imaging systems. The first system PAMIR-Ka is a multichannel pulsed radar system operating at 34 GHz with a very high bandwidth of up to 8 GHz. The second system MIRANDA-94 is a multichannel frequency modulated continuous wave (FMCW) radar with up to 3 GHz at 94 GHz center frequency with a dual polarized antenna. The paper introduces the systems, explains the data collection, and presents first results with respect to multi-look, multi-frequency, multi-polarization, and multi-aspect radar imaging of a test site with military targets.
11:20 IED Command Wire Detection Using Multi-Aspect Processing on SAR Images
Keith T.J. Klein (Delft University of Technology & TNO, The Netherlands); Faruk Uysal (Delft University of Technology, The Netherlands); Miguel Caro Cuenca, Matern Otten and Jacco de Wit (TNO, The Netherlands)
n this paper, a wire detection algorithm is proposed for a circular synthetic aperture radar (SAR) system. The algorithm is specifically designed for SAR images generated from an agile, drone-mounted radar with circular aperture, to be used for the detection of improvised explosive devices (IEDs). A multistage approach consisting of denoising, constant false alarm rate (CFAR) thresholding, feature extraction, and automated detection using the Radon transform, is proposed and applied to a set of SAR images with multiple aspect angles. At each detection step, the look-angles of individual pixels are used to remove false alarms, and improve detection accuracy. The algorithm is tested using measured data and provides an acceptable detection performance on straight wire segments even in the presence of a strong clutter background.
11:40 High- Resolution, Contiguous SAR Imaging Using Co-Located MIMO Arrays: Experimental Proof of Concept
Furkan Korkmaz (The University of Birmingham, United Kingdom (Great Britain) & Necmettin Erbakan University, Turkey); Michail Antoniou (University of Birmingham, United Kingdom (Great Britain))
The paper introduces a novel Synthetic Aperture Radar (SAR) imaging concept with linear Multiple-Input, Multiple-Output (MIMO) radar arrays, which are becoming prolific for short-range sensing in various emerging application spaces. Specifically, the concept makes use of Digital Beam-Forming (DBF) techniques that are enabled in those systems to provide contiguous azimuth imaging, as in stripmap SAR, but with a fine spatial resolution matching that of a spotlight SAR. The main principles of the concept are analytically derived and experimentally verified in laboratory conditions with calibrated targets, as a springboard for a more in-depth system investigation.
12:00 3D Fast Factorized Back-Projection in Cartesian Coordinates
Juliana A Góes (University of Campinas - UNICAMP, Brazil); Valquiria Lima Bessa de Castro (University of Campinas, Brazil); Leonardo Sant´Anna Bins (INPE, Brazil); Hugo Enrique Hernandez-Figueroa (Unicamp, Brazil)
This paper presents a novel 3D Fast Factorized Back-Projection (FFBP) algorithm that is based on an extension of the quadtree approach. A flexible 3D tree structure is generated from a modified Morton order or Z-order curve, a recursive space-filling curve that is suitable for FFBP algorithms. This paper presents, as well, an original method for defining sub-apertures. The proposed algorithm can be applied to any flight path and is about 90 % faster than the direct back-projection, yielding high-resolution 3D images with low phase errors and high degrees of coherence.
12:20 Scanning Radar Target Reconstruction Using Deep Convolutional Neural Network
Jifang Pei (University of Electronic Science and Technology of China (UESTC), China); Deqing Mao (University of Electronic Science and Technology of China, China); Weibo Huo (University of Electronic Science and Technology of China (UESTC), China); Yin Zhang and Yulin Huang (University of Electronic Science and Technology of China, China); Jianyu Yang (School of Electronic Engineering, China)
Target reconstruction is one of the most important missions in the fields of radar signal processing. In this paper, we propose a new deep learning-based approach to reconstruct the target information from the scanning radar returns. Unlike the traditional analytical methods, a deep neural network with a topology of linear chains of convolutional layers is designed, and the input radar signals will be learned layer by layer through the network, which a direct map from the radar echo to the reflectivity function of the targets is obtained during the learning procedure. Finally, we can get the optimal deep learning network as the reconstructing map to recover the scanning radar target information effectively. Simulation results have shown the superiority of the proposed method under different target scenes.

ThM-L8: Radar classification

Room: Ch3
Chairs: Jacques E Cilliers (CSIR, South Africa), Lan Lan (Xidian University, China)
11:00 Target Classification Using Combined YOLO-SVM in High-Resolution Automotive FMCW Radar
Woosuk Kim, Hyunwoong Cho, Jongseok Kim and Byungkwan Kim (Samsung Advanced Institute of Technology (SAIT), Korea (South)); Seongwook Lee (Samsung Advanced Institute of Technology & Machine Learning Lab, Korea (South))
In this paper, we propose a method to classify the human and vehicle objects by combining a support vector machine (SVM) and the deep learning model, you only look once (YOLO), in a high-resolution automotive radar system. To enhance the classification performance, the boundaries of targets estimated from the YOLO model are delivered to the SVM. Then, the overall classification accuracy can be improved by combining the results from the YOLO and SVM with predefined target boundaries. The results show that the proposed method has a better classification performance than those obtained using only YOLO or SVM. From the actual measurements, the proposed method can classify humans and vehicles over 90% accuracy in high-resolution automotive radar.
11:20 Unsupervised Domain Adaptation for Human Activity Recognition in Radar
Xinyu Li (Beijing University of Posts and Telecommunication, China); Xiao jun Jing and Yuan He (Beijing University of Posts and Telecommunications, China)
The difficulty of obtaining large-scale labeled radar data hinders the generalization of supervised deep learning algorithms, which makes the performance of these approaches drop when applied to a new domain. In this paper, we propose an unsupervised adversarial domain adaptation method with radar micro-Doppler spectrograms for human activity recognition. An improved adversarial loss, which has the same mechanism as Generative Adversarial Network, is employed for optimization. Besides, a multi-layer domain discriminator is presented in this method. A simulated dataset (source domain) with labeled radar spectrograms and a measured dataset (target domain) with unlabeled spectrograms are utilized for experiment. Experimental results demonstrate that the proposed method has good performance (with an average accuracy of 81.6%) on recognizing human activities without labeled measured spectrograms in the target domain. And comparison experiments with several state-of-the-art domain adaptation methods also indicate the efficiency of the proposed model.
11:40 A Novel Microwave Staring Correlated Imaging Approach Based on Single Transmitting Antenna
Jianlin Zhang, Bo Yuan, Zheng Jiang, Yuanyue Guo and Dongjin Wang (University of Science and Technology of China, China)
Microwave Staring Correlated Imaging (MSCI) is a high resolution radar imaging technique with its resolution mainly determined by the independence level of radiation fields. In conventional MSCI, an array configuration consisting of multiple transmitting antennas is commonly adopted, which results in high system complexity and implementation difficulty. To remedy these defects, a novel Single Transmitting Antenna MSCI method (STA-MSCI) is proposed in this work. By utilizing frequency-hopping waveforms and variant aperture field patterns in single transmitter scenario, a collection of feasible independent radiation fields is generated and then applied to target reconstruction. Numerical imaging experiments of the proposed method are conducted in a predefined geometric configuration. Simulation results demonstrate that the proposed single transmitting antenna based method achieves superior resolution to the traditional Real Aperture Radar with a simplified system structure, and brings in a new dimension for futher designs of MSCI radiation source.
12:00 Comparison of SEM Methods for Poles Estimation from Scattered Field by Canonical Objects
Yasmina Zaky (Université Cote d'Azur, France); Nicolas Fortino (University of Nice, France); Jean-Yves Dauvignac (Université Côte d'Azur, CNRS, LEAT, France); Fabien Seyfert and Martine Olivi (Inria Sophia Antipolis, France); Laurent Baratchart (INRIA, France)
In this paper, different techniques for SEM poles estimation from the scattered response of an object are explored. Cauchy method and Matrix Pencil are widely used within this field, whereas Vector Fitting method is not often deployed for radar applications. Consequently, we evaluate the accuracy of these techniques applied to the simulated back-scattered response of a PEC sphere and a thin wire for both noisy and noiseless environments, where the precision is based on theoretical poles of each object. To do so, first on PEC sphere, we compute theoretical natural resonant frequencies and set them as references. Then, we apply SEM methods and confront their estimated SEM poles with these reference poles under different SNRs or phase errors. Second, we evaluate these methods on the thin wire by adopting the same approach. Thereby, we conclude about poles estimation precision and robustness of each method when applied to scattered field by canonical objects.
12:20 Feature-Based Classification for Image Segmentation in Automotive Radar Based on Statistical Distribution Analysis
Yang Xiao and Liam Y. Daniel (University of Birmingham, United Kingdom (Great Britain)); Marina S. Gashinova (University of Birmngham, United Kingdom (Great Britain))
Segmentation and potential classification of surface and obstacle regions in automotive radar imagery is the key enabler of effective path planning in autonomous driving. As opposed to traditional radar processing where clutter is considered as an unwanted return and should be effectively removed, autonomous driving requires full scene assessment, where clutter carries necessary information for situational awareness of the autonomous platform and needs to be fully assessed to find the passable areas. In this paper, the statistical distribution features of the radar intensity data of several road-related scenes including asphalt, grass, shadow and target object areas are investigated. The algorithm of classification is developed based on distribution feature extraction and a multivariate Gaussian distribution (MGD) model. Under test dataset recorded by multi-sensor suit was used to evaluate the confusion matrix and F1 score of this classification algorithm.

ThM-SS14: Multisensor multitarget tracking in surveillance applications

Room: Ch1
Chairs: Paolo Braca (CMRE, Italy), Ba-Ngu Vo (Curtin University, Australia)
11:00 GCI Fusion Based Multi-Detection Multitarget Tracking
Lin Gao (University of Florence, Italy); Giorgio Battistelli and Luigi Chisci (Università di Firenze, Italy); Alfonso Farina (Leonardo Company Consultant, Italy)
Multi-detection (MD) observation systems are characterized by multiple observation modes (OMs) and thus simultaneously generate multiple measurements for each target. The main difficulty of exploiting MD systems for multitarget tracking (MTT), in contrast to single-detection (SD) systems, is the great amount of extra computational resources required in order to solve the resulting multidimensional assignment problem. This paper proposes a novel computationally efficient MTT approach for MD systems, wherein a bank of OM-dependent MTT filters with SD model are employed and the OM-dependent posteriors are then fused based on the well-known generalized covariance intersection (GCI) rule. In this way, the computational complexity is significantly reduced compared to existing MTT algorithms with MD model. The effectiveness of the proposed algorithm is assessed by simulation experiments.
11:20 Track Before Detect for Radar Using Stochastic Particle Flow Filters
Fred E Daum and Arjang Noushin (Raytheon, USA)
We have invented a new algorithm that is many orders of magnitude faster than the current state of the art for track before detect problems with radar. We use stochastic particle flow filters to solve this problem for difficult radar applications in dense clutter with closely spaced multiple targets. In some of our applications the density of targets is so high that the radar measurements are often unresolved. Particle filters are particularly good for this application for two reasons: the measurement model is highly nonlinear, and the conditional probability density is extremely multimodal.
11:40 On the Probability of Cross-Radar Assignment Error
Paolo Braca (CMRE, Italy); Leonardo Maria Millefiori (NATO STO CMRE, Italy); Stefano Marano (University of Salerno, Italy); Peter Willett (University of Connecticut, USA); William Dale Blair (Georgia Institute of Technology & Georgia Tech Research Institute, USA)
If two radar sensors observe the same target their measurements can be combined to produce a fused target-state estimate that is of higher quality than that from one radar alone. If there are multiple targets whose information is shared, a necessary first step to fusion is to "assign" each measurement from the first sensor to that at the other in such a way that both refer to the same underlying object, a task generally accomplished by minimizing a global cost involving distance. An assignment error occurs when the measurement originated by target i at the first radar is wrongly associated to a measurement originated by target j (not i) at the second radar. Naturally, when such an error occurs the result is fusion of information describing disparate objects, resulting in degraded estimation performance and poor self-assessment in terms of posterior uncertainty. Here we address the issue, and derive approximate assignment error probability. Remarkably, performance depends only upon the parameters combined to a single scalar constant.
12:00 Progress on the Study for the Use of Long-Range Radars for Space Situational Awareness
Sonia Tomei and Luca Gentile (CNIT RaSS, Italy); Elisa Giusti (CNIT & RaSS, Italy); Marco Martorella (University of Pisa, Italy); Sandro Strappaveccia (CTO Naval Systems, Electronics IT Leonardo spa, Italy); Domenico Vigilante (Leonardo, Italy); Luca Timmoneri (Leonardo Spa, Italy); Alfonso Farina (Leonardo Company Consultant, Italy)
In [1] the authors describe the results achieved in an experimental test conducted with the Leonardo RAT (Radar Avvistamento Terrestre) 31DL long range radar in December 2014 - January 2015 to test the capability of the radar to detect and track Low Earth Orbit (LEO) satellites with an external cue command derived by the NORAD (North American Aerospace Defense Command) Two Line Elements (TLE). After that experimental trials, a research activity was organized by Leonardo spa and RASS CNIT (Radar and Surveillance System - National Interuniversitary Consortium for Telecommunications) with the following main goals: (i) propose new signal and data processing approaches to improve the radar performance and (ii) predict performance obtainable with the network of RAT 31 DL currently deployed on the Italian territory, so demonstrating a dual use capability of radars primarily employed for aerial surveillance. The aim of this paper is to demonstrate the capability in the SST (Space Surveillance and Tracking) scenario of such a network, in particular demonstrating the advantages in terms of coverage maps and accurate orbit determination when using a sensor network. The results are based on simulated data obtained with a simulation tool developed by CNIT RASS, able to analyse different scenarios and orbits.
12:20 Design of NCA Filters for Tracking Maneuvering Targets
William Dale Blair (Georgia Institute of Technology & Georgia Tech Research Institute, USA)
When tracking maneuvering targets with a nearly constant acceleration (NCA) Kalman filter with discrete white noise acceleration (DWNA), the selection of the process noise variance is complicated by the fact that the process noise errors are modeled as white Gaussian, while target maneuvers are deterministic or highly correlated in time. In recent years for nearly constant velocity (NCV) Kalman filters, the deterministic tracking index was introduced and used to develop a relationship between the maximum acceleration of the target and the process noise variance that minimizes the maximum mean squared error (MMSE) in position. A lower bound on the process noise variance was also expressed in terms of the maximum acceleration and deterministic tracking index. More recently, the filter design methods were extended to radar tracking with polar measurements and radar range tracking with LFM waveforms. The design methods for NCV Kalman filters with DWNA were extended to develop design methods for NCV Kalman filters with exponentially-correlated acceleration errors (ECAE) tracking maneuvering targets. In this paper, the design methods for NCV Kalman filters with DWNA are extended to develop design methods for NCA Kalman filters with DWNA tracking maneuvering targets. The effectiveness of the design methods are illustrated via Monte Carlo simulations.

ThM-SS15: Multi-function spectral system co-design

Room: Ch2
Chairs: Daniel W. Bliss (Arizona State University, USA), Athina Petropoulu (Rutgers, USA)
11:00 Joint Radar-Communication-Based Bayesian Predictive Beamforming for Vehicular Networks
Weijie Yuan (University of New South Wales, Australia); Fan Liu and Christos Masouros (University College London, United Kingdom (Great Britain)); Jinhong Yuan and Derrick Wing Kwan Ng (University of New South Wales, Australia)
In this paper, we develop a predictive beamforming scheme based on the dual-functional radar-communication (DFRC) technique, where the road-side units estimates the motion parameters of vehicles exploiting the echoes of the DFRC signals. Compared to the conventional feedback-based beam tracking approaches, the proposed method can reduce the signaling overhead and improve the tracking performance. A novel message passing algorithm is proposed, which yields a near optimal performance achieved by the maximum a posteriori estimation. Simulation results have shown the effectiveness of the proposed DFRC based scheme.
11:20 A Joint Design of MIMO-OFDM Dual-Function Radar Communication System Using Generalized Spatial Modulation
Zhaoyi Xu and Athina Petropulu (Rutgers, The State University of New Jersey, USA); Shunqiao Sun (The University of Alabama, USA)
We propose a novel dual-function radar communication (DFRC) system, consisting of a sparse multiple-input multiple-output (MIMO) radar whose active antennas transmit orthogonal frequency division multiplexing (OFDM) waveforms, and the antenna selection changes between different transmission periods in a controlled manner. The system communicates information via the transmitted OFDM symbols and also the pattern of active transmit antennas in a generalized spatial modulation fashion (GSM). A multi-antenna communication receiver identifies the indices of the active antennas via sparse signal recovery methods. The radar active antennas use most of the OFDM subcarriers in a shared fashion, while each is also allocated a private subcarrier. The sharing of subcarriers among antennas enables high communication bit rate, while the private subcarriers facilitate the construction of a virtual array for a higher angular resolution, and also better estimation on the active antenna indices at the communication receiver. The OFDM waveforms allow the communication receiver to easily mitigate the effect of frequency selective fading. The radar performance of the proposed DFRC system is evaluated via simulations, and bit error rate (BER) results for the communication system are presented.
11:40 Vehicular RF Convergence: Simultaneous Radar, Communications, and PNT for Urban Air Mobility and Automotive Applications
Andrew Herschfelt, Alex Chiriyath and Daniel W. Bliss (Arizona State University, USA); Christ D. Richmond (Arizona State University & Ira A. Fulton Schools of Engineering, USA); Urbashi Mitra (University of Southern California, USA); Shannon D Blunt (University of Kansas, USA)
Modern RF environments are becoming increasingly congested. This limits the opportunities and capabilities of modern RF systems, obstructing the development and proliferation of new technologies. Novel vehicular RF technologies promise a new era of transportation capabilities, but legacy design techniques cannot adapt to the current spectral congestion. We summarize recent RF Convergence results and discuss how they mitigate spectral congestion in modern vehicular applications. We propose a joint radar, communications, positioning, navigation, and timing (JRCPNT) system architecture as a suitable candidate for future automotive applications. We present relevant performance bounds and initial experimental results to demonstrate the potential performance enhancements of such multiple-function RF systems. We define multiple-channel, multiple-user receiver (MCMUR) techniques that enable this architecture and discuss how these components cooperate to enable these performance enhancements. We summarize initial experimental results to demonstrate the viability of such multiple-function architectures in the context of urban air mobility (UAM) and other automotive applications.
12:00 Multicarrier DS-CDMA Waveforms for Joint Radar-Communication System
Salil Sharma, Maarit Melvasalo and Visa Koivunen (Aalto University, Finland)
Multicarrier waveforms have always been of high importance in many radar and communication applications as they have many desirable properties. Multicarrier Direct Sequence-CDMA (MC-DS-CDMA) is a well known modulation approach. It can be adapted to joint radar-communication systems with appropriate design choices. For the radar sensing purposes it is desirable to have ambiguity function resembling the ideal thumbtack function. In this paper we propose a joint radar-communication waveform stemming from MC-DS-CDMA. A subset of P subcarriers are dedicated to communications using orthogonal variable spreading factor codes and a scrambling code to facilitate different rates and reduced correlation among subcarrier signals. Radar subsystem uses spreading codes for ranging with desirable auto- and cross-correlation properties and a scrambling code to separate the radar and communication signals. It is demonstrated that the proposed waveforms have highly desirable ambiguity functions (AF). We study the peak-to-average power ratios of the waveforms, too.
12:20 Hybrid Beamforming for Multi-User Dual-Function MIMO Radar-Communication System
Ziyang Cheng (University of Electronic Science and Technology of China, China); Shengnan Shi (UESTC, China); Bin Liao (Shenzhen University, China)
This paper investigates a hybrid beamforming design for multi-user dual-function radar-communication (DFRC) system. The hybrid transmit/receive beamformers are designed by maximizing the sum-rate under constraints of power as well as similarity between the designed beamformer and the reference one that has good beampattern property. To solve the nonconvex optimization problem, we devise a two-stage method. To be more specific, in the first stage, we obtain the fully digital beamformer based on the weighted mean-square error minimization (WMMSE) method, in the second stage, we iteratively optimize the digital and analog beamformers to approximate the achieved fully digital beamformer. Numerical simulations are provided to demonstrate the effectiveness of the proposed schemes.

ThM-SS16: Ground based radar remote sensing of clouds and precipitation

Room: Ch4
Chairs: Luca Baldini (Consiglio Nazionale delle Ricerche, Italy), Frank S. Marzano (Sapienza University of Rome, Italy)
11:00 Development and Observations of the Phased Array Radar for Weather Application
Tomoo Ushio (Osaka University, Japan); Eiichi Yoshikawa (JAXA, Japan); Hiroshi Kikuchi (The University of Electro-Communications, Japan); Masakazu Wada (Toshiba Corporation, Japan)
Recent progress of information and communication technologies has been enabling us to realize a rapid scanning radar system. In 2012, Toshiba Corporation and Osaka University succeeded in developing a new type of Phased Array Radar (PAR) system (Yoshikawa et al. 2013, Ushio et al. 2015) under a grant of National Institute of Communication and Information Technology (NICT), and installed in Suita Campus, Osaka University. This PAR system can scan the whole sky within 30 seconds up to 60 km in radius over 100 elevation angles with digital beam forming technique, and the initial observation results demonstrate the unique capability of the new PAR system. After its installation, a new clutter mitigation algorithm from adaptive beam forming technique was developed and tested with the PAR system, and succeeded in suppressing not only the ground clutter but also ghost echo from strong precipitation echo nearby more than 20dB. And also the adaptive algorithm was applied to supress the range sidelobe and showed the sidelobe level of -60dB. Upon this success, a new experiment started in 2015 under a grant of SIP (Strategic Innovation Promotion Program) to create the Osaka Urban Phased Array Radar Demonstration Network. The main sensor of the Osaka Network is a 2-node Phased Array Radar Network and lightning location system. In this experiment, data products including reflectivities, VIL (Vertically Integrated Liquid water content), precipitation rates and others are transferred to Osaka Local Government in real time to prevent water related natural hazards. In this presentation, system architecture and some results are presented.
11:20 Triple Frequency Doppler Spectra Closure Study Across the Melting Layer: Lessons Learnt
Kamil Mroz (UK-NCEO, United Kingdom (Great Britain)); Alessandro Battaglia (Politecnico of Turin, Italy & University of Leicester, United Kingdom (Great Britain)); Stefan Kneifel, Leonie von Terzi, Markus Karrer and Davide Ori (University of Cologne, United Kingdom (Great Britain))
By exploiting novel measurements from vertically pointing multi-frequency Doppler radars, this study investigates the link between rain and ice microphysics across the melting layer in stratiform rain, the main source of precipitation in the mid-latitudes. A closure study is proposed in this paper where Drop Size Distributions (DSD) are retrieved from multi-frequency radar Doppler spectra and propagated upward to predict the Particle Size Distributions (PSD) in the overlying snow. Snow PSDs are retrieved above the freezing level for several ice models from the full Doppler spectra measured at such level. The model that best matches the measured triple frequency spectra is used to infer the PSDs, which are then compared to the PSDs predicted from rain from a melting-only steady-state assumption. Overall, the predicted and the retrieved mean mass weighted diameters (D_m) of ice are highly correlated, with the retrieved D_m being on average circa 15% larger. Although, a correlation between the precipitation rates above and below the melting zone is weaker (CC = 0.66), the melted equivalent accumulation over 6 h period is nearly perfectly matched (1% difference only). This novel methodology can be applied to assess the validity of some assumptions like the constant precipitation rate across the melting region via long-term observations; this will advance our knowledge of the processes occurring across the melting region. This has the potential to improve space-borne radar precipitation retrievals as well, especially those from the Dual Frequency Precipitation Radar on-board of the Global Precipitation Measurement core satellite.
11:40 AI in Weather Radars
V. Chandrasekar (Colorado State University, USA)
Modern ground-based weather radars are mostly dual-polarized and they are rich in information content in multiple dimensions, and are ideal candidates for effective artificial intelligence applications. There are a large number of dual-polarization radars around the world. In addition, space borne weather radars such as Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar (PR) and Global Precipitation Measurement (GPM) mission Dual-frequency Precipitation Radar (DPR) have also produced rich observations, that are great to be analyzed using AI. AI has already been used with weather radar information long before it came popular in mainstream, such as use of neural networks for ingesting vertical profiles and using neuro-fuzzy systems for hydrometeor classification. The primary goal of this paper is to familiarize three major applications areas for Weather radar AI namely, Nowcasting, precipitation System Classification and Quantitative precipitation Estimation. These three applications domains have engaged very different types of AI technologies.
12:00 Nonlinear Iterative Optimization Approaches for Pulse Compression in Weather Radars
Luca Facheris and Fabrizio Argenti (University of Florence, Italy)
The problem of pulse compression in weather radars is particularly challenging due to the distributed nature of the targets and to the high gradients of reflectivity that can be encountered along range. This poses strong requirements in terms of low energy of the sidelobes and low peaks of the output of the receive filter, in order to retrieve correctly the range profiles of reflectivity. The pulse compression approaches presented in this paper are based both on a matched filter and on a mismatched receiver. The pursued objective is minimum sidelobe energy and peak level through a strategy based on nonlinear iterative optimization.

Thursday, September 24 12:40 - 2:00 (Europe/Rome)

WIE Panel: Gender bias in the workplace

Chair: Silvia L Ullo (Università degli Studi del Sannio, Italy)

Thursday, September 24 2:00 - 3:00 (Europe/Rome)

Plenary Talk 3 - Mahta Moghaddam (University of Southern California, USA): Microwave Sensing for Medical Imaging and Monitoring of Thermal Therapies: Accelerated Inverse Scattering via Learning.

Chair: Moeness G. Amin (Villanova University, USA)

Thursday, September 24 3:00 - 4:40 (Europe/Rome)

ThA-L1: Radar tracking

Room: Ch1
Chairs: William Dale Blair (Georgia Institute of Technology & Georgia Tech Research Institute, USA), Luigi Chisci (Università di Firenze, Italy)
3:00 PLA-JPDA for Indoor Multi-Person Tracking Using IR-UWB Radars
Hongyu Qian, Xiuzhu Yang, Xinyue Zhang, Yi Ding and Lin Zhang (Beijing University of Posts and Telecommunications, China)
Impulse-Radio Ultra-WideBand (IR-UWB) radar is of great interest due to its high range resolution. In indoor environments, IR-UWB radars are used for multi-person tracking. Besides, the Joint Probabilistic Data Association (JPDA) has been designed to associate tracks with persons. However, when multiple persons are wandering close to each other, multipath clutter and track coalescence pose challenges to the tracking task. This paper proposes the Path-Loss-based Adaptive JPDA (PLA-JPDA) for indoor multi-person tracking using IR-UWB radars. The PLA-JPDA suppresses multipath clutter according to the distance-dependent path-loss of radar signals. Moreover, the PLA-JPDA adjusts the validation gate based on measurements and modifies the weights related to association probability. Experiments are conducted in an indoor environment with three IR-UWB radars tracking two persons to validate the proposed approach. In the experiments, two persons wander towards each other along parallel lines. The average root-mean-square-error (RMSE) of tracking with the proposed PLA-JPDA is 0.14 meters. It is 18.7% and 13.7% of the average RMSEs of tracking with the JPDA and the Nearest Neighbor JPDA (NN-JPDA), respectively. The PLA-JPDA also outperforms the JPDA and the NN-JPDA in terms of Optimal Sub-Pattern Assignment (OSPA) distance. The source codes and experimental data of this paper are published at https://github.com/qhybupt/PLA-JPDA.
3:20 A Track-Before-Detect Approach for UWB Radar Sensor Networks
Bo Yan, Andrea Giorgetti and Enrico Paolini (University of Bologna, Italy)
Networks of ultra-wideband (UWB) radar nodes have been shown to represent a practical and effective approach for precise localization and tracking of moving non-collaborative persons and objects. In UWB radar sensor networks (RSNs), the presence of strong clutter, weak target echoes, maneuvering targets, closely spaced targets, and extended targets are obstacles to achieve a satisfactory tracking performance. In this work, a track-before-detect (TBD) approach for UWB RSNs is proposed. Both spatial information and temporal relationship between measurements are exploited in generating several possible candidate trajectories, from which only the more suitable are selected. Experimental results carried out with actual UWB signals in the presence of a human target confirm the effectiveness of TBD paradigm in UWB RSNs. Remarkably, the localization error obtained by the proposed RSN can be within 0.4 m with no false trajectories, and no misdetection despite the presence of heavy clutter and the low signal-to-noise ratio (SNR) regime.
3:40 Radar Fusion for Multipath Mitigation in Indoor Environments
Evert Ismael Pocoma Copa (Universite Libre de Bruxelles, Belgium); Kheireddine Aziz (Interuniversity Microelectronics Centre (IMEC) & Vrije Universiteit Brussel (VUB), Belgium); Maxim Rykunov, Eddy De Greef and Andre Bourdoux (IMEC, Belgium); François Horlin (Université libre de Bruxelles, Belgium)
One of the main challenges of radar-based localization applications in indoor environments is the presence of strong multipath. When the radar bandwidth is large enough, multipath components can be resolved in range but they result in unwanted ghost targets. We propose a novel multipath mitigation approach that exploits the fact that multipaths are highly dependent on the scene geometry. The multipath mitigation approach discards the ghost targets based on the fused information of multiple radars located at different positions in the scene. For such radar fusion, the output of the radar signal processing chain is translated into the world coordinate system that is common for all the radars. We propose a radar alignment approach to estimate the translation and rotation parameters from radar to world coordinate system and vice versa. Our multipath mitigation method is combined with an unscented Kalman filter to improve the localization accuracy. We demonstrate the effectiveness of our complete approach with a real experiment using two radars to detect and track a target in a room with severe multipath.
4:00 BAAS: Bayesian Tracking and Fusion Assisted Object Annotation of Radar Sensor Data for Artificial Intelligence Application
Stefan Haag and Bharanidhar Duraisamy (Daimler AG, Germany); Felix Govaers (Fraunhofer FKIE / University of Bonn, Germany); Wolfgang Koch (Fraunhofer FKIE & University of Bonn, Germany); Martin Fritzsche and Juergen Dickmann (Daimler AG, Germany)
This paper introduces BAAS, a new Extended Object Tracking (EOT) and fusion-based label annotation framework for radar detections in autonomous driving. Our framework utilizes Bayesian-based tracking, smoothing and eventually fusion methods to provide veritable and precise object trajectories along with shape estimation to provide annotation labels on the detection level under various supervision levels. Simultaneously, the framework provides evaluation of tracking performance and label annotation. If manually labeled data is available, each processing module can be analyzed independently or combined with other modules to enable closed-loop continuous improvements. The framework performance is evaluated in a challenging urban real-world scenario in terms of tracking performance and the label annotation errors. We demonstrate the functionality of the proposed approach for varying dynamic objects and class types.
4:20 Note on Autocorrelation of the Residuals of the NCV Kalman Filter Tracking a Maneuvering Target
Paul Miceli (Georgia Tech Research Institute, USA); William Dale Blair (Georgia Institute of Technology & Georgia Tech Research Institute, USA)
The Kalman filter is the optimal estimator for tracking a nearly constant velocity (NCV) target with white noise acceleration errors. However, when the target maneuvers the white noise input assumption is violated and the filter estimate will become biased or worse, lead to the loss of the target track. Selection of a larger process noise will reduce the bias in the state estimates during a maneuver, but when the target is not maneuvering the filter performance is far from optimal. In decision-directed techniques for tracking maneuvering targets, the estimation process is impacted by hard decisions regarding onset and termination of target maneuvers. In this work, the autocorrelation of consecutive residuals for a maneuvering target is derived and a new test for maneuver detection is proposed. When the process noise covariance of an NCV Kalman filter is artificially high for tracking maneuvering targets, achieving a specified false rate for maneuver detection is a challenge. A remedy to this challenge is also presented.

ThA-L2: Software-defined, small radar architectures, prototypes

Room: Ch2
Chairs: Braham Himed (AFRL, USA), Pierfrancesco Lombardo (University of Rome La Sapienza, Italy)
3:00 Reviewing the Application and Integration of Software Defined Radios to Radar Systems
Shiqi Feng (Per Vices Corporation, Canada); Neelam Mughees (COMSATS University Islamabad, Pakistan); Victor Wollesen (Per Vices Corporation, Canada)
The advent of software-defined radio (SDR) is driving the traditional radar systems to reach new heights where true versatility can be achieved cost-effectively with careful and thorough integration. Thanks to the latest advances in semiconductor manufacturing capabilities, data converter technologies, digital signal processing techniques, and software designs, SDR has made itself from a novel technology to the cornerstone of next-generation radio platforms with endless possibilities. One of the most important motivations behind the incorporation of SDR into radar design is its ability to quickly adapt to future emerging applications while meeting today's performance and reliability requirements by moving most of the challenges from costly hardware redesign to the software stack. To accomplish this, however, sufficient design margin in conjunction with meticulous planning and rigorous integration effort must be put in place. This article reviews how SDR has gradually become the backbone of modern radar architecture along with emphasis on illustrating the benefit, limitations, and challenges associated with the integration process. We also note potential solutions associated with the various integration challenges and describe the required analysis designers should carry out to better realize the benefits of SDR.
3:20 SDR-Based Hardware Implementation and Performance Measurement of Transmit Beampattern Design Algorithms
Ahmad Bin Rabiah and Mohammed Alsakabi (Prince Sultan Defense Studies and Research Center, Saudi Arabia); Omar Aldayel and Saleh A Alshebeili (King Saud University, Saudi Arabia)
The utilization of multiple transmit antenna elements in radar systems can enhance the efficiency of energy consumption, detection probability and mitigation of clutter and interference. Recently, many analytical methods have been developed to exploit this by a proper design of the transmit array to achieve the desired beampattern. However, the hardware implementation of these methods is challenging due to some practical issues such as mutual coupling, the coherency requirements for the excitation of transmit waveforms, and the non-linearity of high power amplifiers. In this paper, we implement three of the state-of-art waveform design methods in an array transmit configuration, present the experimental results of their true measured transmit beampattern, and compare it to the simulated results. Our experimental hardware consists of Commercial Off-The-Shelf (COTS) equipment along with an anechoic chamber to absorb multi-path reflections.
3:40 The UAV Radar Imaging Prototype Developed in the Frame of the VESTA Project
Giuseppe Esposito (IREA-CNR, Italy); Carlo Noviello (IREA-CNR & University of Napoli Federico II, Italy); Francesco Soldovieri (CNR, Italy); Giancarmine Fasano (University of Naples Federico II, Italy); Graziano Gagliarde (TopView srl, Italy); Gianluca Luisi and Federico Saccoccio (TopView Srl, Italy); Ilaria Catapano (IREA-CNR, Italy)
This work presents an UAV based radar imaging prototype designed in the frame of the VESTA project, an Italian regional project funded in the frame of the POR Campania FESR 2014/2020 program. The key elements of the prototype can be summarized as follows: 1) a radar payload, which is designed according to the payload constrains due to the use of a small UAV platform and is realized by using hardware components available on the market; 2) a Carrier based Differential Global Positioning system, which is used to track the UAV trajectory with a centimetric accuracy; 3) an ad hoc designed data processing strategy based on a microwave tomographic approach enhanced by a motion compensation procedure, capable of providing high resolution images in the vertical image domain, i.e., the plane defined by the along-track and the nadir directions. The effectiveness of the prototype is assessed thanks to an experimental test regarding targets of different size and electromagnetic properties.
4:00 An All-COTS High Sampling Frequency Pulse-Doppler Imaging Radar
Russell H Kenney (University of Oklahoma & Advanced Radar Research Center, USA); Kurt Konyalioglu, Mark Yeary and Hjalti Sigmarsson (University of Oklahoma, USA); Jay W McDaniel (University of Oklahoma & Advanced Radar Research Center, USA)
This paper presents a high sampling frequency pulse-Doppler radar that was constructed using commercially available components. The system architecture is explained with emphasis placed on the digital transceiver subsystem. A calibration technique is presented to mitigate range ambiguity due to digital latency variation between resets of the system. The pulse-Doppler operation of the system is then evaluated through the range-Doppler mapping of moving and stationary targets, confirming the radar's potential for synthetic aperture radar imaging.
4:20 Experimental Implementation of a Multi-Antenna 802.11Ax-Based Passive Bistatic Radar for Human Target Detection
Laurent Storrer (Univesité Libre de Bruxelles (ULB), Belgium); Hasan Can Yildirim and Evert Pocoma (Université Libre de Bruxelles (ULB), Belgium); Jerome Louveaux (Université catholique de Louvain, Belgium); Philippe De Doncker (ULB, Belgium); Sofie Pollin (KU Leuven, Belgium); François Horlin (Université libre de Bruxelles, Belgium)
We investigate and experimentally demonstrate a multi-antenna Wi-Fi-based passive bistatic radar (PBR) to perform indoor range-Doppler-angle detection of human targets. The latest Wi-Fi standard, 802.11ax, is considered as signal of opportunity, enabling a high range resolution suited for indoor detection. We build a Uniform-Linear-Array (ULA) using Universal Software Radio Peripherals (USRPs) as PBR receiver (RX), and present a novel calibration method to compensate the hardware-induced phase shift difference between the signals from the different antennas of the ULA. To avoid data association problems and limitations on the number of detectable targets for the Direction-of-Arrival (DoA) estimation, we demonstrate theoretically the possibility to use only the cell of the target in the radar range-Doppler maps (RDMs) across antennas as input to the Multiple Signal Classification (MUSIC) algorithm, rather than using the raw received signals. We validate the experimental setup and the processing by detecting the range, speed and DoA of two human targets moving in a room.

ThA-SS17: Radar for health monitoring and biomedical applications

Room: Ch3
Chairs: Fauzia Ahmad (Temple University, USA), Francesco Fioranelli (TU Delft, The Netherlands)
3:00 Joint Waveform/Receiver Design for Vital-Sign Detection in Signal-Dependent Interference
Gabriel Beltrão (University of Luxembourg & Interdisciplinary Centre for Security, Reliability and Trust, Luxembourg); Mohammad Alaee-Kerahroodi (Interdisciplinary Center for Security, Reliability and Trust, Université du Luxembourg, Luxembourg); Udo Schröder (IEE S.A., Luxembourg); Bhavani Shankar Mysore R (Interdisciplinary Centre for Security, Reliability and Trust & University of Luxembourg, Luxembourg)
This paper presents the joint design of discrete slow-time radar waveform and receive filter, with the aim of enhancing the Signal to Interference and Noise Ratio (SINR) in phase coded radar systems for vital-sign monitoring. Towards this, we consider maximizing the SINR at the input of the vital-sign estimation block, when transmitting hardware efficient M-ary Phase Shift Keying (MPSK) sequences. This multi-variable and non-convex optimization problem is efficiently solved based on a Minimum Variance Distortionless Response (MVDR) filter, with the Coordinate Descent (CD) approach for the sequence optimization, and the obtained results have shown attractive interference suppression capabilities, even for the simple binary case.
3:20 Comparative Study of Radar Architectures for Human Vital Signs Measurement
Etienne Antide (Université Grenoble Alpes & CEA, LETI, France); Mykhailo Zarudniev (Université Grenoble Alpes, France); Olivier Michel (INPG, France); Michael Pelissier (CEA-Leti/Minatec, France)
Radars can be used as a non-invasive solution to monitor the vital signs of patients. The heart and respiratory rates are generally extracted by analyzing the phase variations of the radar signal, thus motivating the use of millimeter-waves. This, however, comes at the cost of a higher attenuation with the distance of travel, which in turn lowers the signal to noise ratio. While the state-of-the-art considers various architectures of millimeter-wave (mmW) radar for vital signs extraction, they are seldom compared in terms of hardware complexity and power consumption even though these aspects are of utmost importance for autonomous applications. This paper presents a comparative analysis of the state-of-the-art short-range radar solutions and takes into account their respective hardware complexity needed to improve the signal to noise ratio. It aims to select the most relevant low-power architecture for an autonomous application. Analytical models for power estimations are presented and compared with simulation results. Finally, in light of these performances the architectures complexities over their respective hardware limitations are discussed.
3:40 Distributed Radar Information Fusion for Gait Recognition and Fall Detection
Haobo Li and Julien Le Kernec (University of Glasgow, United Kingdom (Great Britain)); Ajay Mehul (The University of Alabama, USA); Sevgi Z Zubeyde Gurbuz (University of Alabama & TUBITAK Space Technologies Research Institute, Italy); Francesco Fioranelli (TU Delft, The Netherlands)
This paper discusses a fusion framework with data from multiple, distributed radar sensors based on conventional classifiers, and transfer learning with pre-trained deep networks. The application considered is the classification of gait styles and the detection of critical accidents such as falls. The data were collected from a network comprised of one Ancortek frequency modulated continuous wave radar and three ultra wide-band Xethru radars. The radar systems within the network were placed in three different locations, notably, in front of participants, on the ceiling, and on the right-hand side of the monitored area. The proposed information fusion framework compares feature level fusion, soft fusion with the classifier confidence level, and hard fusion with Naïve Bayes combiner (NBC). Regarding the classifier, linear SVM, Random-Forest Bagging Trees, and five pre-trained neural networks are introduced to the fusion algorithm, where the VGG-16 network yields the best performance (about 84%) with the help of NBC. Compared to the best cases with conventional classifiers, it is reported that 20% and 16% subsequent improvement are achieved for individual usage of single radar and fusion.
4:00 Body Movement Cancellation Using Adaptive Filtering Technology for Radar-Based Vital Sign Monitoring
Li Zhang and Chuanwei Ding (Nanjing University of Science and Technology, China); Xudong Zhou (Nanjing University of Science and Technology,China); Hong Hong (Nanjing University of Science and Technology, China); Changzhi Li (Texas Tech University, USA); Xiaohua Zhu (Nanjing University of Science and Technology, China)
Noise due to body movement is a critical challenge for noncontact vital signs detection. This paper presents an adaptive filter-based body movement cancellation solution that does not require any hardware modification to a conventional continuous-wave (CW) Doppler biomedical radar. When a body movement is detected, the recently detected segments without body movement is used as the reference to process the segment corrupted by body movement with an adaptive filter. The result suggests that the proposed method can reduce the impact of body movement on vital signs monitoring.
4:20 An Unsupervised Approach for Graph-based Robust Clustering of Human Gait Signatures
Aylin Tastan (Technische Universität Darmstadt, Germany); Michael Muma and Abdelhak M Zoubir (Darmstadt University of Technology, Germany)
Classification of gait abnormalities plays a key role in medical diagnosis, sports, physiotherapy and rehabilitation. We demonstrate the effectiveness of a new graph construction-based outlier detection method and and the applicability of a new parameter-free clustering approach on radar-based human gait signatures. Micro-Doppler radar-based human gait signatures of ten test subjects for five different gait types consisting of normal, simulated abnormal and assisted walks are clustered using five different clustering algorithms. The proposed algorithm outperforms existing methods both in cluster enumeration and partition and achieves an overall correct clustering rate of 92.8%. The developed method is promising for performing medical diagnosis in a robust unsupervised fashion.

ThA-SS18: Satellite sensing of the atmosphere: radar technologies and methods for advancing atmospheric and climate science.

Room: Ch4
Chairs: Luca Facheris (University of Florence, Italy), Simone Tanelli (Jet Propulsion Laboratory, California Institute of Technology, USA)
3:00 Spaceborne Atmospheric Radar Technology Development
Lihu Li, Gerald Heymsfield and Matthew McLinden (National Aeronautics and Space Administration, USA); Paul Racette (NASA GSFC, USA); Michael Cooley, Peter Stenger and Thomas Spence (Northrop Grumman Corporation, USA)
Spaceborne atmospheric radar technologies have been developed at the NASA Goddard Space Flight Center (GSFC) in collaboration with Northrop Grumman Corporation (NGC). This paper gives an overview of future spaceborne radar concepts and key technologies targeted for large satellite platforms, the International Space Station (ISS), and SmallSat applications.
3:20 A Compact W-band Breadboard Radar for Atmospheric Measurements
Raquel Monje, Robert Beauchamp, Ken Cooper, Shivani Joshi and Simone Tanelli (Jet Propulsion Laboratory, California Institute of Technology, USA)
We present the W-band channel breadboard and first-light measurements of a new atmospheric profiling radar, called CloudCube. CloudCube is intended to be a multifrequency, ultra-compact, low-cost, modularized radar for vertical profiling of clouds, convection and precipitation structures and dynamics from space. The instrument uses a radar architecture, where the baseband signal is directly upconverted to the RF band without any addition of intermediate frequencies or further multiplication scheme. This architecture, combined with pulse compression techniques, provides a simple and effective solution while achieving the required performance and robustness of an airborne or space flight radar instrument.
3:40 Cloud Dynamics Revealed by a G-band Humidity-Sounding Differential Absorption Radar
Ken Cooper (Jet Propulsion Laboratory, California Institute of Technology); Robert Beauchamp, Richard Roy, Luis Millan, Matt Lebsock and Raquel Monje (Jet Propulsion Laboratory, California Institute of Technology, USA)
VIPR, Vapor Inside-cloud Profiling Radar, is a low peak power cloud radar operating over 167-174.8 GHz, which is on the flank of the 183 GHz water vapor absorption line. VIPR's primary purpose is to measure absolute humidity by using the frequency dependence of water vapor absorption along the radar beam's path. Here we describe an additional capability of VIPR, which is to reveal internal cloud dynamics from the power ratio of range compressed spectra from long-duration linear-frequency-modulated pulses with opposite chirp directions. While the measurements are qualitative, relying on differences of incoherently averaged cloud reflectivity that arise from the Doppler-shift induced range migration of spatial reflectivity gradients, we show that they can reveal the location of wind shear in a cloud system.
4:00 RainCube and Its Legacy for the Next Generation of Spaceborne Cloud and Precipitation Radars
Simone Tanelli (Jet Propulsion Laboratory, California Institute of Technology, USA); Eva Peral, Eastwood Im and Mauricio Sanchez-Barbetty (Jet Propulsion Laboratory, USA); Robert Beauchamp and Raquel Monje (Jet Propulsion Laboratory, California Institute of Technology, USA)
The multidimensional challenge of observing cloud and precipitation processes from space has motivated and focused a number of technological developments in the last decade. They span from novel compact instrument architectures to modular millimeter wave phased array solutions. In this paper we summarize the main capabilities offered by three instruments from this new generation, and discuss how they affect formulation of future observing systems.
4:20 Ground-Based Radar Insight into Warm Marine Boundary Layer Clouds for Shaping Future Spaceborne Radar Missions
Katia Lamer (Brookhaven National Laboratory Stony Brook, USA); Pavlos Kollias (Stony Brook University, United Kingdom (Great Britain)); Alessandro Battaglia (Politecnico of Turin, Italy & University of Leicester, United Kingdom (Great Britain)); Simon Preval (University of Leicester, United Kingdom (Great Britain))
This study relies on ground-based observations to quantify the limitations of current spaceborne radars when it comes to capturing the coverage and vertical structure of Warm Marine Boundary Layer (WMBL) clouds. Its finding that neither the CloudSat-Cloud Profiling Radar (CPR) nor the EarthCARE-CPR will be able to accurately characterize cloud cover, cloud fraction and the location of a large fraction of WMBL clouds further encourages the development of concepts for new spaceborne radar systems targeting WMBL clouds. Given the properties of WMBL clouds extracted from ground-based radar measurements collected in the eastern northern Atlantic, we recommend that the next generation of space-borne radars targeting WMBL science shall operate interlaced pulse modes including both a highly sensitive long-pulse and a less sensitive but clutter limiting short-pulse mode.

ThA-SS19: Millimeter-wave synthetic aperture radar

Room: Ch5
Chairs: Kumar Vijay Mishra (United States Army Research Laboratory, USA), Lam Nguyen (Army Research Laboratory, USA)
3:00 k-Space Decomposition Based Super-resolution Three-dimensional Imaging Method for Millimeter Wave Radar
Tomoki Omori (University of Electro-Communications, Japan); Yusuke Isono, Katsuhiko Kondo and Yusuke Akamine (SOKEN, Japan); Shouhei Kidera (University of Electro-Communications, Japan)
Millimeter wave imaging radar is indefensible for collision avoidance of self-driving system, especially in optically blurred visions.The range points migration (RPM) is one of the most promising imaging algorithms, which provides a number of advantages from synthetic aperture radar (SAR), in terms of accuracy, computational complexity, and potential for multi-functional imaging. The inherent problem in the RPM is that it suffers from lower angular resolution in narrower frequency band even if higher frequency e.g. millimeter wave, signal is exploited. To address this problem, the k-space decomposition based RPM has been developed. This paper focuses on the experimental validation of this method using the X-band or millimeter wave radar system, and demonstrated that our method significantly enhances the reconstruction accuracy in three-dimensional images for the two simple spheres and realistic vehicle targets.
3:20 High-Throughput 3-D Millimeter-Wave Imaging of Packaged Goods
Andreas Pedross-Engel (University of Washington & ThruWave Inc., USA); Claire Watts and Matthew Reynolds (University of Washington, USA)
This work presents a first of its kind high-throughput 3-D millimeter-wave imaging system for packaged goods in commerce. We use millimeter waves to image through non-metallic packaging to see the items inside, and permit automated analysis of crucial business and process metrics such as object count, anomaly detection, or estimation of void space inside packages. The presented system achieves system resolutions of δx = 20.4 mm, δy = 4.3 mm, and δz = 8.4 mm, with accuracy of 1 mm in the y direction and 2.5 mm in the z direction. It is shown that the system is not only capable of producing 3-D images of the products inside an optically-opaque corrugated cardboard shipping box, but is also sensitive enough to produce an image of the cardboard itself.
3:40 Near-Field MIMO-ISAR Millimeter-Wave Imaging
Josiah W. Smith (University of Texas at Dallas, USA); Muhammet Yanik and Murat Torlak (The University of Texas at Dallas, USA)
Multiple-input-multiple-output (MIMO) millimeter-wave (mmWave) sensors for synthetic aperture radar (SAR) and inverse SAR (ISAR) address the fundamental challenges of cost-effectiveness and scalability inherent to near-field imaging. In this paper, near-field MIMO-ISAR mmWave imaging systems are discussed and developed. The rotational ISAR (R-ISAR) regime investigated in this paper requires rotating the target at a constant radial distance from the transceiver and scanning the transceiver along a vertical track. Using a 77GHz mmWave radar, a high resolution three-dimensional (3-D) image can be reconstructed from this two-dimensional scanning taking into account the spherical near-field wavefront. While prior work in literature consists of single-input-single-output circular synthetic aperture radar (SISO-CSAR) algorithms or computationally sluggish MIMO-CSAR image reconstruction algorithms, this paper proposes a novel algorithm for efficient MIMO 3-D holographic imaging and details the design of a MIMO R-ISAR imaging system. The proposed algorithm applies a multistatic-to-monostatic phase compensation to the R-ISAR regime allowing for use of highly efficient monostatic algorithms. We demonstrate the algorithm's performance in real-world imaging scenarios on a prototyped MIMO R-ISAR platform. Our fully integrated system, consisting of an mechanical scanner and efficient imaging algorithm, is capable of pairing the scanning efficiency of the MIMO regime with the computational efficiency of single pixel image reconstruction algorithms.
4:00 Image Enhancement with Blind Deconvolution in Millimeter-Wave 3-D FLoSAR
Kumar Vijay Mishra (United States Army Research Laboratory, USA); Lam Nguyen (Army Research Laboratory, USA)
Recently, three-dimensional (3-D) millimeter-wave (mm-Wave) Forward-Looking Synthetic Aperture Radar (FLoSAR) has emerged as a promising technology for applications such as self-landing, navigation, and collision-avoidance in a deteriorated vision environment. The FLoSAR images targets when the radar platform is moving in the forward direction. The advantage of operating at mm-Wave comes from potentially exploiting its unlicensed, wide bandwidth for high-resolution imaging. Simultaneously, high angular resolution is difficult to achieve in FLoSAR because objects located symmetrically along the flight direction have the same range and Doppler history. Additionally, at mm-Wave, GPS accuracy is insufficient to obtain platform position at subwavelength level leading to challenges in providing motion compensation in the reconstructed images. In this work, we investigate the autofocusing by modeling it as a sparse deconvolution problem. The radar measurements correspond to a convolution of unknown radar image that is convolved with an impulse function with unknown focusing errors. In particular, we aim to exploit the sparse nature of the target scenario and compare the results with the phase gradient autofocus method.
4:20 A Review of Recent Advancements Including Machine Learning on Synthetic Aperture Radar Using Millimeter-Wave Radar
Arindam Sengupta (The University of Arizona, USA); Feng Jin, Reydesel A Cuevas and Siyang Cao (University of Arizona, USA)
In this paper, we review recent and emerging Synthetic Aperture Radar (SAR) applications using mm-Wave radar, ranging from concealed item detection to autonomous systems. Furthermore, relevant machine learning (ML) concepts are introduced and the review of ML applications in high-resolution mm- Wave SAR image enhancement and generation are presented. The paper is concluded with challenges and expectations of mm- Wave SAR imaging with emphasis on autonomous vehicles.

Thursday, September 24 4:40 - 5:00 (Europe/Rome)

Coffee Break

Thursday, September 24 5:00 - 5:20 (Europe/Rome)

ThA-D1: DEMO 1

Room: Ch1
5:00 Audio Sonar for Gaining Hands-On Experience of Radar Principles
Warren P du Plessis (University of Pretoria, South Africa)
While learning the theory underlying radar is important, it is also necessary to gain hands-on experience to truly understand the relevant concepts. However, radar systems are expensive and often perform important functions making it unlikely that students will be able to experiment with them. Furthermore, large spaces and expensive hardware are required as a result of eclipsing in pulsed radars. An audio sonar will be demonstrated to show how important concepts such as ambiguities, clutter, and others can be illustrated using a simple, low-cost audio sonar.

ThA-D2: DEMO 2

Room: Ch2
5:00 Radar for Structural Monitoring and Assets Mapping
Fabio Giannino and Guido Manacorda (IDS GeoRadar, Italy); Alessandro Simi (IDS GeoRadar, Greece); Matteo Cecchetti (IDS GeoRadar srl, Italy); Daniele Vacca (IDS GeoRadar, Italy)
The use of radars in Civil Engineering is well known and there are several applications where it is currently utilised; they include the location of buried objects with Ground Penetrating Radar (GPR) equipment and, more recently, the real-time monitoring of buildings and structures' stability with Interferometric Ground-Based SAR (GB-InSAR) measurements.

ThA-D3: DEMO 3

Room: Ch3
5:00 Hardware Demonstration of a Scalable Cognitive Sparse Array
Satish Mulleti and Yariv Shavit (Weizmann Institute of Science, Israel); Moshe Namer (Technion - Israel Institute of Technology, Israel); Yonina C. Eldar (Weizmann Institute of Science, Israel)
High-resolution direction of arrival estimation requires a large number of antenna elements which increases the computational cost, hardware complexity, and power requirements. To balance between hardware complexity and resolution, recently, we proposed a cognitive, scalable, sparse array selection technique based on a submodular-greedy algorithm. In this demo, we present a design and implementation of a hardware prototype that demonstrate the proposed sparse antenna selection strategy. Through real-time experiments, we show that the proposed sparse selection method results in a 2-3 dB lower error compared to a typically employed random selection method.

Thursday, September 24 5:20 - 6:40 (Europe/Rome)

ThA-P1: PSS3 - Multitarget tracking

Chairs: Giorgio Battistelli (Università di Firenze, Italy), Suqi Li (University of Electronic Science and Technology of China, China)
Distributed Multi-Target Tracking over an Asynchronous Multi-Sensor Network
Guchong Li (University of Electronic Science and Technology of China, China); Giorgio Battistelli and Luigi Chisci (Università di Firenze, Italy); Lingjiang Kong (University of Electronic Science and Technology of China, China)
This paper addresses distributed multi-target tracking (DMTT) over an asynchronous multi-sensor network (AMSN). Within the AMSN, the sensor nodes are usually misaligned in time due to different sampling instants and/or rates. At the same time, time-offsets among nodes are always imprecise or even unknown. In such cases, time alignment (TA) needs to be carried out before fusion of information between different nodes. In the considered AMSN for DMTT, Probability Hypothesis Density (PHD) filters are run in each node for propagating in time a local first-order statistic, called intensity, of the target set, while arithmetic average (AA) fusion is used to combine intensities from different nodes. Recalling that AA intensity fusion admits an information-theoretic interpretation in terms of minimizer of the weighted average of Cauchy-Schwartz divergences (CSDs) with respect to the local intensities, the corresponding minimum weighted average CSD (MWCSD) is adopted as cost to be minimized for TA purposes. To ensure good convergence of the TA parameters, a convex combination of the instantaneous cost and the squared difference between current and previous estimates, is proposed. Furthermore, a sampling technique is adopted to solve the optimization problem. Finally, simulation experiments are provided to demonstrate the effectiveness of the proposed approach.
A Decentralized PHD Filter for Multi-Target Tracking in Asynchronous Multi-Static Radar System
Qi Yang and Wei Yi (University of Electronic Science and Technology of China, China); Mahendra Mallick (Independent Consultant, USA)
Asynchronism among the local radars is one of the most important challenges for multi-target tracking (MTT) in multi-static radar systems. In order to address this challenge in the framework of random finite set (RFS) based on Bayesian inference, we propose a decentralized probability hypothesis density (PHD) filter based on the asynchronous periodical sequential estimation (APSE). First, starting from a multi-target Bayesian filter, we derive the multi-target density update expressions for the APSE solution in the RFS framework. Next, we develop the PHD recursion expressions of the APSE solution, named as APSE-PHD, and describe the Gaussian mixture (GM) implementation of the APSE-PHD. Simulation results for a challenging tracking scenario confirm that the proposed APSE-PHD algorithm is effective for MTT in the asynchronous multistatic radar system and outperforms the existing PHD-based algorithm.
An Efficient PHD Filter for Multi-target Tracking with Out-of-Sequence Measurement
Qi Yang and Wei Yi (University of Electronic Science and Technology of China, China)
In this paper, we address the update problem of the out-of-sequence measurement (OOSM) in multi-sensor multitarget tracking (MTT). Based on the probability hypothesis density (PHD) filter in the framework of random finite set (RFS), we propose an efficient algorithm for OOSM update in the MTT scenario, named as E-OOSM-PHD, which involves two stages: retrodiction and posterior update. The stage of retrodiction facilitates the incorporation of the OOSM at the appropriate time. The posterior update is to incorporate the OOSM and compensate the cardinality at the current time, so as to improve the multi-target tracking performance. The effectiveness of the proposed algorithm is demonstrated in a challenging tracking scenario via simulation results.
Distributed Multi-object Tracking and Registration with LMB Filter in Multistatic Radar Systems
Lei Chai, Wei Yi, Wujun Li and Lingjiang Kong (University of Electronic Science and Technology of China, China)
This paper addresses the problem of distributed multi-object tracking (MOT) in a multistatic radar system wherein each radar has no knowledge about the positions of its neighboring nodes in its local coordinate system. The labeled multi-Bernoulli filters (LMB) are running locally in each radar for MOT. The Weighted Arithmetic Average (WAA) fusion rule is employed due to its cardinality robustness and computational efficiency. In our method, after obtaining the local LMB multi-object densities (MOD), the radar positions are estimated by minimizing the Cauchy-Schwartz (CS) divergence between the matched Bernoulli densities. In addition, we propose a new WAA fusion method for LMB MODs, the WAA fusion is also performed between the matched Bernoulli densities, then the fused LMB MOD is reconstructed by the fused Bernoulli densities. Hence the joint multi-object tracking and registration problem can be effectively solved. The efficacy of the proposed method is demonstrated in a challenging tracking scenario via numerical experiments.
Resource Allocation for Multi-Target Tracking in Multi-Static Radar Systems with Imperfect Detection Performance
Xiujuan Lu, Jun Sun, Ye Yuan and Xiaobo Yang (University of Electronic Science and Technology of China, China)
In this paper, by considering an imperfect detection environment, a joint node selection and power allocation (JNPA) strategy for multi-static radar (MSR) system with the task of multi-target tracking (MTT) is proposed. Current resource allocation works are often studied based on an ideal detection precondition that the sensors detect targets with the probability of detection (P_D) equals to 1. It is practically impossible to achieve due to the influence of signal radiation attenuation and fluctuating of target amplitudes. To address this drawback, the proposed JNPA takes the practical case where P_D ≤ 1 into consideration. The centralized fusion framework is adopted by the MSR to obtain the global MTT results. Accordingly, the posterior Cramer-Rao lower bound (PCRLB) with P_D ≤ 1 is derived and utilized as the tracking performance metric of single target. Then, an overall cost function (OCF) is established to describe the global MTT performance. By utilizing the OCF as the objective function and combining with system resource constraints, the JNPA is formulated as an optimization problem. The optimization problem is non-convex. Next, we propose a two-step method to solve it. Simulation results verify the effectiveness of the proposed JNPA strategy.

ThA-P2: PSS4 - Multisensor multitarget tracking

Chair: Nicola Forti (NATO STO CMRE, Italy)
A Non-Markovian Prediction for the GM-PHD Filter Based on Recurrent Neural Networks
Isabel Schlangen, Steffen Jung and Alexander Charlish (Fraunhofer FKIE, Germany)
Bayesian multi-target filtering has become an essential signal processing technique for a plethora of applications, most prominently for radar, sonar, and image processing. It provides an automated way to study the dynamics of objects based on a set of carefully chosen process and sensor models. However, the estimation performance strongly depends upon the suitability of those models and a poor match between the true object behaviour and its describing model can lead to grave misinterpretations of the situation, especially in the presence of ambiguities or missed detections. Traditionally, analytic solutions such as the Near-Constant Velocity (NCV) or the Constant Turn (CT) models are selected to describe the target dynamics, often under the Markov assumption that disregards information before the current time step. In this paper, a Gaussian Mixture Probability Hypothesis Density (GM-PHD) filter is presented whose transition function is a Recurrent Neural Network with Long Short-Term Memory that is especially designed to predict a full Gaussian density. It is shown in simulation that the proposed method prevents the filter from overestimating the number of objects in the presence of false alarms or missed detections and helps resolve ambiguities when targets are close to each other.
Random Finite Set Tracking for Anomaly Detection in the Presence of Clutter
Nicola Forti and Leonardo Maria Millefiori (NATO STO CMRE, Italy); Paolo Braca (CMRE, Italy); Peter Willett (University of Connecticut, USA)
In this paper, a sequential Bayesian framework is proposed to address the task of joint anomaly detection and tracking for surveillance applications in the presence of clutter. This is achieved by modeling the anomaly as a switching unknown control input which goes into action by modifying the expected dynamics of a target and ceases its activity (becomes nonexistent) under nominal behavior. Random Finite Sets (RFS) make it possible to represent the switching nature of the object anomalous behavior and derive a hybrid Bernoulli filter (HBF) that sequentially updates the joint posterior density of a Bernoulli RFS for the unknown velocity input and the object kinematic state. In addition, the proposed HBF has been customized for maritime anomaly detection by using a piecewise Ornstein-Uhlenbeck (OU) stochastic process as dynamic model of vessels. We illustrate the effectiveness of the proposed filter, implemented in Gaussian-mixture form, and compare its performance in a maritime surveillance example with the Interacting Multiple Model Probabilistic Data Association Filter (IMM-PDAF) for different levels of clutter.
Multiple-Hypothesis Group Tracking
Stefano Coraluppi and Constantino Rago (Systems and Technology Research, USA); Craig Carthel (Systems & Technology Research, USA)
This paper generalizes the track-oriented multiple-hypothesis formalism to consider a time-varying number of groups, each composed of a time-varying number of targets. Additionally, the paper develops a tractable multi-stage tracking approach that performs target-level tracking under suitable statistics in the first stage, followed by group-level tracking under the generalized recursion in the second stage. Our approach is relevant to group tracking problems, with the restriction that objects do not change from one group to another. It is also relevant to repeated-measurement problems for extended targets.

ThA-P3: PSS5 - Radar networks for climate change

Chairs: Frank Gekat (Leonardo Germany, Germany), Stefano Turso (Fraunhofer Institute for High Frequency Physics and Radar Techniques FHR, Germany)
Weather Detection with an AESA-Based Airborne Sense and Avoid Radar
Florian Kunstmann, Dietmar Klarer, Anouk Puchinger and Stefan Beer (Hensoldt Sensors GmbH, Germany)
Modern radar systems shall provide multifunctional capabilities because airborne applications typically demand sensor equipment, which has optimized size, weight and power. Therefore the use of multifunctional sensors is necessary and efficient. This work represents the first weather radar data of flight tests of such an multifunctional AESA (Active Electronically Scanned Array) based combined Weather and Sense and Avoid (S&A) Radar for airborne application in UAVs (Unmanned Aerial Vehicles) of class HALE/MALE (High/Medium Altitude Long Endurance). To achieve feasible results, appropriate clutter suppression is needed. The outcome is compared with ground weather data from the Deutscher Wetterdienst.
Application of Micro Rain Radar for Supporting Quantitative Weather Radar Precipitation Measurements
Gerhard Peters and Piet Markmann (METEK GmbH, Germany); Hans-Jürgen Kirtzel (Metek GmbH, Germany)
The "Micro Rain Radar" is a vertically pointing Doppler radar operating as a remote sensor for drop size distributions. A concept is proposed to use this technique to support quantitative precipitation retrieval by weather radar.
A Novel Antenna Concept for Weather Applications Based on a Cylindrical Parabolic Reflector
Stefano Turso (Fraunhofer Institute for High Frequency Physics and Radar Techniques FHR, Germany); Carlos Salzburg (Fraunhofer, Germany); Marc Vizcarro (Hensoldt, Germany); Thomas Bertuch (Fraunhofer FHR, Germany)
Wide adoption of phased array weather radars based on large active panels has been limited so far by an intrinsic cost increase respect to bi-axial mechanically steered solutions. A novel antenna concept is proposed to achieve a more cost effective solution by decoupling gain generation and electronic steering. A simple parabolic cylinder with active feed for electronic steering in elevation, mechanically rotated in azimuth, is shown to yield acceptable degradation of the cross-polarization performance of the feed.
Validation of Space-Borne Radar Precipitation Retrievals Using Disdrometer Data over Italy
Elisa Adirosi (Italian National Research Council (CNR), Italy); Luca Baldini (Consiglio Nazionale delle Ricerche, Italy); Mario Montopoli (ISAC CNR, Italy); Alessandro Bracci (University of Bologna, Italy)
In spite of the high relevance of the satellite data for collecting information regarding precipitation at global scale, validating satellite retrievals with measurements collected by sensors at ground is still a challenging task. To date, the Dual-frequency Precipitation Radar (DPR) aboard the Core Satellite of the Global Precipitation Measurement (GPM) mission is the only one able to provide, at global scale, vertical profiles of rainfall rate, radar reflectivity, and Drop Size Distribution (DSD) parameters. In this study, we compare near surface GPM retrievals with almost 6 years of DSD parameters estimated by a laser disdrometer in Rome (Italy) since the launch of GPM mission. The comparison shows limited difference in performances of the different GPM algorithms, also with reference to the dual frequency algorithm with respect to the single frequency one. Furthermore, the agreement between satellite and ground based data depends on the considered precipitation variable as well as on the spatial and temporal scale matching constrains.
Regional Precipitation Mosaicking Using Multifrequency Weather Radar Network in Complex Orography
Errico Picciotti (HIMET, Italy); Stefano Barbieri (CETEMPS, University of L'Aquila, Italy); Saverio Di Fabio (CETEMPS, Italy); Klaide De Sanctis (HIMET, Italy); Raffaele Lidori (University of L'Aquila, Italy); Frank S. Marzano (Sapienza University of Rome, Italy)
Networks of weather radars are ideally suited to provide remotely sensed data of the state of the atmosphere over a wide geographic area providing updated views for severe weather detection and near real-time forecasting. Nonetheless measurements made by weather radars contain a tremendous amount of information, yet considerable effort should be made to extract out scientifically and operationally meaningful products. This work shows recent advancements in terms of structure, processing and mosaicking methods specifically implemented for the local radar network covering the Abruzzo Region, in central Italy, whose hydrological risks are further enhanced by its complex orography. All algorithms have been developed by using Python packages that provides a variety of routines for reading, processing, analyzing, compositing and visualizing data from different weather radar systems.

ThA-P4: Indoor, GPR and through-the-wall radars

Chair: Francesco Soldovieri (CNR, Italy)
Accurate People Counting Based on Radar: Deep Learning Approach
Jae-Ho Choi, Ji-Eun Kim and Nam-Hoon Jeong (POSTECH, Korea (South)); Seung-Hyun Jin (Defense Agency for Technology and Quality, Korea (South)); Kyung-Tae Kim (Pohang University of Science and Technology (POSTECH), Korea (South))
In this study, a novel radar-based people counting (PC) method is presented using the deep learning (DL) approach. The DL algorithm is a great tool that enables the automatic formation of the optimal features; however, it must be utilized carefully, considering the domain knowledge to prevent the concerns of learning unnecessary information, followed by overfitting. To address the problem and successfully apply the DL framework to the radar-based PC, we propose three novel solutions. First, we establish the preprocessing pipelines to transform the raw signals into a suitable form for network inputs. Second, a network architecture is newly proposed considering the radar signal characteristics and PC application. Finally, we propose several data augmentation strategies to artificially increase the size of training data. It was observed from experiments using real measured data that the proposed DL-based PC approach outperforms the conventional PC methods.
Super-resolution Doppler Velocity and Range Estimator for Short-range Human Recognition Radar
Takeru Ando and Shouhei Kidera (University of Electro-Communications, Japan)
Super-resolution range and Doppler velocity estimation scheme based on novel algorithm is presented in this paper, assuming a human recognition using microwave or millimeter wave radar. Micro-Doppler analysis for human recognition requires some breakthroughs to overcome a limitation of temporal and Doppler velocity resolutions. As a promising method for the above issue, the weighted kernel density (WKD) estimator has been proposed. However, the WKD still faces the problem that it hardly decomposes each reflection response from each part of human body, in particular within range resolution. To address with this issue, this paper newly incorporates the k-space decomposition scheme into the WKD method. The numerical test, using the simplified 3-D human body model, demonstrate that the proposed method remarkably enhances the accuracy for range and Doppler velocity estimations.
NLP Based Skeletal Pose Estimation Using mmWave Radar Point-Cloud: A Simulation Approach
Arindam Sengupta (The University of Arizona, USA); Feng Jin and Siyang Cao (University of Arizona, USA)
Human skeletal pose estimation can find several applications ranging from remote patient monitoring, pedestrian detection to defense security and surveillance. However, traditionally used high-resolution vision based sensors suffer operationally during poor illumination or object occlusion. Radars can overcome these challenges, albeit at the cost of a lower resolution. mmWave radars, on account of a higher bandwidth, have the ability to represent a target as a sparse point-cloud, with a higher resolution than its traditional radar counterparts. A supervised learning approach is adopted for skeletal estimation from the point-cloud, as its random nature from frame-to-frame makes explicit point-to-point association non-trivial. However, the lack of available radar data-sets make it extremely difficult to develop machine-learning aided methods to improve radar based computer vision applications. In this paper, we present a study to use simulated mmWave-radar-like point-cloud data to estimate skeletal key-points, of a human target using, a natural language processing approach. The sparsity and randomness in the radar point-cloud is simulated from a Microsoft Kinect acquired data using a random sampling approach. Two consecutive frames of the simulated radar point-cloud are first voxelized and aggregated, and a seq2seq architecture is used for "summarizing" it to the desired skeletal keypoints. Simulated data obtained by randomly sampling from a combination of (i) corrupting the 3- D ground truth skeletal coordinates with Gaussian noise over a range of varying degrees of variance, and (ii) adding random point-cloud noise to the corrupted data, is used to evaluate the performance of the model. The comprehensive methodology, results and discussion is presented in this paper. The promising results from this proof-of-concept simulation study serve as a basis for future experimental study using mmWave radars which will also be made open-access for public research and development of radar based perception and computer-vision.
Improved People Counting Algorithm for Indoor Environments Using 60 GHz FMCW Radar
Jonas Weiß (TU Munich, Germany); Rodrigo Pérez and Erwin Biebl (Technische Universität München, Germany)
This work presents a novel people counting system based on an adaptive peak detection and a vital sign verification algorithm with a 60 GHz frequency modulated continuous wave radar sensor, which uses the antenna in package technology. Thus, the sensor has a small size, a low energy consumption, and a low price. The algorithm applies an adaptive ordered statistic constant false alarm rate detection on the filtered and preprocessed range spectrum to detect human targets, and checks them in a second step with a vital sign verification algorithm. In the experimental setup a dataset of 0 to 4 people present in an office room was recorded with three sensors from different viewing angles. A plus-minus-one accuracy of up to 85.4 % was achieved and the ability to generalize to new sensor positions is shown.
Analyzing the Classification Capability of Micro-Doppler Spectra
Hans-Günter Hirsch (Niederrhein University of Applied Sciences, Germany); Jan Stähler (Institute for Pattern Recognition, Germany); Manfred Hägelen and Reinhard Kulke (IMST GmbH, Germany)
The Micro-Doppler spectra of a frequency modulated continuous wave (FMCW) radar system can be used to identify objects in a surveillance application scenario. We investigated different classification approaches to determine their potential to classify different objects. A radar sensor in the 24 GHz band was used to record scenes with three different moving objects. We compared classification methods which recognized fairly short segments of the recorded scene to other approaches which analyzed the whole sequence of Doppler spectra with a Gaussian Mixture - Hidden Markov modelling (GMM-HMM) and with a long short-term memory (LSTM) based neural network. High recognition rates of up to 99% could be achieved on the recorded test data set.
Radar Based Deep Learning Technology for Loudspeaker Faults Detection and Classification
Alessio Izzo and Carmine Clemente (University of Strathclyde, United Kingdom (Great Britain)); Ludovico Ausiello (Solent University, United Kingdom (Great Britain)); John J Soraghan (University of Strathclyde, United Kingdom (Great Britain))
Recently, radar based micro-Doppler signature analysis has been successfully applied in various sectors including both defence and civilian applications. A joint radar micro-Doppler and deep learning technology for End-Of-Line (EOL) test of loudspeakers is proposed in this paper. This approach offers the potential benefits of characterizing the mechanical motion of a loudspeaker in a noisy environment as a production line, in order to automatically identify and classify defects. Starting from real radar signal, the proposed Bidirectional Long Short-Term Memory (BiLSTM) classifier has been tested on training, validation and test dataset. The results show that the proposed approach produces a probability of correct classification above the 98%, outperforming the traditional k-NN classifier.
Detailed Feature Representation and Analysis of Low Frequency UWB Radar Range Profile for Improving Through-wall Human Activity Recognition
Yansong Peng and Shisheng Guo (University of Electronic Science and Technology of China, China)
Human Activity Recognition (HAR) by using low frequency Ultra-wideband (UWB) through-the-wall radar (TWR) has always been a challenging task. Due to the limited bandwidth of signal, one can obtain few information for human activity from the common range profile. In this paper, we proposed a novel preprocessing method based on a kind of pixel-changing fringe pattern for the original range profile, which can present more detailed features. The effectiveness of the proposed method was validated by the simulated and experimental data. Meanwhile, the output recognition results of nine human activities for classic convolutional neural network (CNN) shown that the range profile data processed by the proposed method is more suitable for HAR.
Combat Gesture Classification Using Through-The-Wall Radar Based on Multi-domain Features Association
Beili Ma and Baixiao Chen (Xidian University, China)
Based on the method of multi-domain micro-Doppler features association, classification of hand gesture activities is considered and studied in this paper. We first focus on the differences of several typical individual combat gestures by using the through-the-wall radar to collect human posture and motion information. A database of real measured radar data with more than 2000 recordings from 3 different human subjects has been collected in a series of experiments. Based on the micro-Doppler signature, six key features are extracted through the data analysis in time domain, frequency domain and time-frequency domain respectively. And the obtained feature vectors are used alone or in combination to input to the random forest classifier for training. The experiment results show that the classification accuracy is found to be more than 90%.
A Ground Penetrating Radar Data Reconstruction Method Based on Generation Networks
Longhao Xie (University of Electronic Science and Technology, China); Qing Zhao and Jianjian Huo (University of Electronic Science and Technology of China, China); Guo Cheng (University of Electronic Science and Technology of China, Italy)
In this paper, a ground penetrating radar (GPR) data reconstruction method based on generation networks is proposed. The main purpose of this method is to reconstruct the GPR data from degraded data such as missing traces data and sparse sampling data. The generation networks can obtain the reconstructed GPR data by training the network mapping degraded data from a two-dimensional random sequence. It can be used for obtaining denser GPR data and recovering missing GPR traces. Both simulated and field data are used to illustrate the validity. It could still be well reconstructed after 50% of the traces were removed.
Health Assessment of Trees Using GPR-derived Root Density Maps
Livia Lantini, Fabio Tosti, Iraklis Giannakis, Lilong Zou and Daniel Egyir (University of West London, United Kingdom (Great Britain)); Dale Mortimer (Tree Service, London Borough of Ealing, United Kingdom (Great Britain)); Amir M. Alani (University of West London, United Kingdom (Great Britain))
In the current paper, a coherent framework for estimating the density and the distribution of roots using ground penetrating radar is presented. The proposed methodology is a multi-stage data processing scheme that is applied in semi-circular measurements collected concentrically around the investigated tree. The employed processing pipeline consists of three distinct and sequential steps. In the first step, the raw B-scans are subject to time-zero correction, zero-offset removal, time-varying gain and the singular value decomposition (SVD) filter. The SVD filter is used in order to effectively eliminate multiples and ringing noise from the B-Scans and increase the overall signal-to-clutter ratio. The second step consists of a tracking algorithm that aims at identifying patterns that resemble tree roots. In the last step, a continuous function is fitted to each root in order to effectively interpolate between points and subsequently estimate the density of the roots. The current paper concludes with a case study on a diseased urban tree at the Gunnersbury Park, London, United Kingdom. The investigated area has been excavated after the measurements took place in order to accurately validate and assess the performance of the suggested processing pipeline.
Knowledge-assisted Building Layout Reconstruction for Through-the-wall Radar Imaging
Yang Zhang (University of Electronic Science and Technology of China, China); Guolong Cui (University of Electronic Science and Technology of China (UESTC), China); Qingxin Ran, Huquan Li, Shisheng Guo and Jiahui Chen (University of Electronic Science and Technology of China, China)
In this paper, the reconstruction of building layout and interior object is considered only utilizing the attenuation power of transmitted narrow-band signal. More specifically, bistatic radars scan the entire unknown area in the form of tomographic imaging to complete the reconstruction issue. Based on the idea of knowledge assistance, we propose a priori information constrained algebraic reconstruction technology (PIC-ART) algorithm, which presents better performance than current stateof- the-art algorithms in multi-media area. Besides, we first utilize the electromagnetic (EM) simulation software to verify the feasibility of the established model under omnidirectional antennas. The EM simulation results have validated the effectiveness of the proposed algorithm.
A Real-Time Human Activity Recognition Method for Through-the-Wall Radar
Can Cheng and Fei Ling (University of Electronic Science and Tech, China); Shisheng Guo (University of Electronic Science and Technology of China, China); Guolong Cui (University of Electronic Science and Technology of China (UESTC), China); Qiang Jian, Chao Jia and Qingxin Ran (University of Electronic Science and Tech, China)
Human activity recognition (HAR) has long been a research hotspot in anti-terrorism, patient monitoring and other applications. Throughout the current research progress, the realtime recognition of unknown start and end time has not been well solved. In addition, the recognition accuracy of blocking human in the wall-through scene needs to be improved. To tackle this issue, this paper proposed a novel range profile sequence driven end-to-end model, which specifically employed random cropping training method. Then, we carried out experiments on actual Through-the-Wall Radar(TWR), finally achieved an average accuracy of 97.6% on four common activities, and the most significant is that our method can immediately output recognition results , without waiting for the end of activity.

ThA-P5: ECCM, defense and technology

Chair: Luca Timmoneri (Leonardo Spa, Italy)
Nodes Positions Optimization for Sparse Large Aperture Radar System with Enhanced Target Search Efficiency
Minglong Deng, Ziyang Cheng and Chunchun Zheng (University of Electronic Science and Technology of China, China); Zi-Shu He (University of Electronics Science and Technology of China, China)
This paper focuses on the problem of the optimization of nodes positions for sparse large aperture radar (SLAR) system. To overcome the deficiency of low target search efficiency in traditional SLAR system, a search scheme based on grating lobe controlling, in which the SLAR system searches multiple directions simultaneously by utilizing a transmit beampattern with multiple grating lobes, is provided. More specifically, the relationship between the nodes positions and grating lobes is mathematically discussed. On this basis, under the practical node position constraint, a design method of nodes positions for synthesizing a beampattern with desired grating lobes and low sidelobes is presented. Some representative simulations are provided to validate the effectiveness of the proposed scheme.
Novel Radar North Correction Estimation Algorithm
Ondrej Nemec (University of Pardubice, Czech Republic); Jan Pidanic (University of Pardubice & Faculty of Electrical Engineering and Informatics, Czech Republic); Pavel Sedivy (Retia, a. s., Czech Republic)
The paper is dedicated to the novel software radar north correction estimation algorithms based on received terrain and ground target echoes, vector map data and height data. The current implementation of algorithms utilizes OpenStreetMap and SRTM height data. The developed algorithms can offer an efficient alternative to currently used north-seeking methods and devices. Algorithms were tested on radar data of short-range 2D radar ReVISOR produced by RETIA a.s company. North-seeker algorithm results were validated for various sites with comparable accuracy to HW north seekers.
A Study on the Memory Effect of Mixer
Bowei Zhang, Ling Tong and Bo Gao (University of Electronic Science and Technology of China, China); Baoguo Yang, Shubiao Li, Shengli Liang and Fushun Nian (Science and Technology on Electronic Test & Measurement Laboratory, China)
As we know, with the increase of input signal bandwidth, the asymmetry of upper and lower sideband of the third-order intermodulation distortion (IMD) will become more evident. The industry describes this asymmetric phenomenon as the memory effect of the microwave component. For simplicity, the third-order intermodulation distortion (IM3), the major part of IMD, is chosen in investigating the conditions for IMD asymmetry. With the rapid communication systems, the bandwidth of the input signal becomes wider and wider, and the memory effect becomes very significant. Therefore, it's necessary to study the asymmetry of IM3. Today, taking the advantages of three signal generators and a spectrum analyzer, the first investigation into the effect of different excitation conditions to the memory effect of double balanced diode mixer. The different excitation conditions include frequency separation, the power of the radio-frequency (RF) signal and the local oscillator (LO) signal, and so on. For the first time, the measurement results show the impact of the input signal (LO signal and RF signal) power and the relative size of LO frequency and RF frequency on mixer memory effect.
Determination of Radiometric Radar Cross-Section
Massimiliano Rossi (MBDA Italia S.p.A., Italy); Marco Frasca (MBDA Italia, Italy)
In this paper we show as, similarly to active radars, a concept of radar cross-section applies also for the case where the sensor is a radiometer in the microwave range. In this way, a complete theoretical framework to design such kind of passive device can be given for their use as passive radars.
Single Data Set Side-lobe Cancellation Method for Non-stationary Clutter Suppression in HF Mixed Mode Surface Wave Radar
Jiazhi Zhang, Xin Zhang, Weibo Deng, Yang Qiang and Mengxiao Zhao (Harbin Institute of Technology, China)
High-frequency mixed mode surface wave radar is a new kind of over-the-horizon radar system and is capable of early warning and marine exclusive economic zone monitoring. Due to combination of the ionospheric channel and the sea surface, the inhomogeneous background makes the clutter statistics change rapidly in even adjacent range cells. This sample starved scenario degrades the clutter suppression performance for conventional adaptive filter methods which utilize the training data from neighboring range cells. In this paper, a novel single data set side-lobe cancellation method is proposed to overcome this challenge. Firstly, the non-stationary property of the first-order sea clutter using measured data is analyzed. Then, the clutter suppression algorithm, which is called SDS-SLC, is presented. Finally, the validity of the proposed method is verified by the experimental data and the comparison result shows it outperforms other traditional methods.
Multi-Receiver Deception Jamming Against SAR Using Frequency Diverse Array
Jianfei Yu, Mu Zhou and Zengshan Tian (Chongqing University of Posts and Telecommunications, China); Bang Huang (University of Electronic Science and Technology of China, China)
This paper proposes a multi-receiver deception jamming approach against the synthetic aperture radar (SAR) using the frequency diverse array (FDA). Different from the traditional phased-array antennas, the FDA utilizes a slight frequency offset across elements, and thus multiple false targets are potentially generated in the range dimension of the SAR image. Moreover, utilizing the instantaneous slant distance difference between different receivers, the jammer can directly reduce the reliance on parameters determined by the SAR working mode such as the velocity and the squint angle of the SAR, which are hard to be obtained by the electronic reconnaissance. Such a new method can generate multiple predesigned false targets or scenes without knowing geometric parameters of the SAR. The effectiveness of the proposed method is validated by the experimental simulation results.
Empirical Calibration Coefficient Estimation in the Context of Multi-Functional RF Systems
Michael Kohler, Daniel W O'Hagan and Josef Worms (Fraunhofer FHR, Germany); Oliver Bringmann (University of Tuebingen, Embedded Systems / FZI, Germany)
This paper presents an empirical approach to estimate the transfer function required for the calibration in general of a two-channel multi-functional broadband receiver developed at Fraunhofer FHR. Modern broadband receivers may enable increased functionality, provided they maintain adequate performance. For example, some modern active electronically scanned arrays (AESAs) and, increasingly, future systems may incorporate simultaneous operating modes, particularly multiple different radar modes together with communications and spectrum sensing modes. It is clear, therefore, that broadband hardware calibration is an essential feature for the realization of Multi-Functional RF Systems (MFRFS). In this paper the characteristics of the MFRF receiver are described via its transfer function. Under laboratory conditions, the receiver transfer function may be determined classically using a vector network analyzer. However, for in-theatre applications, on-site calibration will be necessary. A defined input signal is fed into the broadband receiver via an external signal source and the output signal is recorded. The receiver transfer function is then determined by comparing the input and output signals. A coherent signal processing approach using empirical transfer function estimation and a corresponding measurement concept will be presented resulting in reduced measurement and computational effort compared to state-of-the-art. The performance of the proposed approach is evaluated using real measurement data from a laboratory setup.
Performance Analysis and Optimal Placement for Unambiguous DOA Estimation Using Interferometry
Huaiyuan Liang and Xiangrong Wang (Beihang University, China); Alfonso Farina (Leonardo Company Consultant, Italy)
Interferometry is considered as a computationally efficient and economically fair method for direction of arrival (DOA) estimation using a limited number of receiving antennas. Moreover, interferometric arrays are capable of providing relatively high-accuracy estimates, with the sacrifice of involving phase ambiguity as a byproduct of large antenna spacing. In this work, we first propose a phase line length matching (PLLM) method for unambiguous DOA estimation. Then, we analyze estimation performance of three-antenna interferometric arrays in both cases of high and low signal-to-noise ratio (SNR). In the case of low SNR, dominant estimation error comes from phase line mismatching, and a metric of line mismatching probability (LMP) is proposed to quantify the performance. In the case of high SNR, a closed-form formula of conditional mean squared error (CMSE) for DOA estimation is derived and compared with Cramer-Rao Bound (CRB). Finally, the influence of interferometric array configuration on estimation performance is discussed and optimal configurations in both cases are suggested as well. Comparative simulation results validate the effectiveness of proposed PLLM method and the superiority of optimal interferometric arrays for unambiguous DOA estimation.
High Resolution Harmonic Radar ISAR Experiments and System Design Strategies
Florian Bischeltsrieder, Andreas Heinzel and Markus Peichl (German Aerospace Center (DLR), Germany)
Harmonic radar is a promising radar technique for the detection of electronic devices in security related applications. In contrast to established and commercially available systems, where the radar works in conjunction with a known and cooperative transponder responding at a harmonic frequency, the goal here is to detect and localize non-cooperative targets such as mobile phones. In this work, a combination of classic and harmonic radar in one system is presented. The system layout together with the design strategies and experiences gained is described in detail. Measurement results indicate how an inverse synthetic aperture radar (ISAR) setup could be beneficial in a real world scenario.
Exclusion of Target Signals from Training Samples for Robust Adaptive Beamforming in Radar
Ryuhei Takahashi (MitsubishiElectricCorporation, Japan); Toshio Wakayama (Mitsubishi Electric Corporation, Japan)
A technique for excluding the target signal from contaminated training samples is proposed for robust adaptive beamforming (ABF) in radar. We focus on applying the proposed technique to mitigate the effect of ionospheric clutter in high-frequency surface-wave radar systems, but the technique can also be applied to overcome jamming signals without the need to allocate dedicated listening intervals. A computer simulation indicates that the proposed technique for effectively excluding the target signal is successful; the subsequent ABF achieves asymptotically the same signal to interference-plus-noise ratio where target contamination in the training samples is absent.
A Random Antenna Subset Selection Jamming Method Against Multistatic Radar System
Xiangtuan Wang, Yimin Liu and Tianyao Huang (Tsinghua University, China)
Multistatic radar system (MSRS) is considered an effective scheme to suppress mainlobe jamming, since it has higher spatial resolution enabling jamming cancellation from spatial domain. To develop electronic countermeasures against MSRS, a random array subset selection (RASS) jamming method is proposed in this paper. In the RASS jammer, elements of the array antenna are activated randomly, leading to stable mainlobe and random sidelobes, different from the traditional jammer that applies the complete antenna array enjoying constant mainlobe and sidelobes. We study the covariance matrix of jamming signals received by radars, and derive its rank, revealing that the covariance matrix is of full rank. We also calculate the output jamming to signal and noise ratio (JSNR) after the subspace-based jamming suppression methods used in MSRS under the proposed jamming method, which demonstrates that the full rank property invalidates such suppression methods. Numerical results verify our analytical deduction and exhibit the improved countermeasure performance of our proposed RASS jamming method compared to the traditional one.
Clustering for Jamming Environment Classification
Vincenzo Carotenuto (University of Naples Federico II, Italy); Antonio De Maio (University of Naples "Federico II", Italy); Salvatore Iommelli (Ente di Formazione Professionale Maxwell, Italy)
A hierarchical clustering architecture is proposed to deal with the problem of jamming environment classification when multiple noise-like jammers are possibly present. Assuming the availability of clutter free multichannel data, a two-level hierarchical procedure is devised to unveil the presence of clusters containing range cells experiencing the same jamming interference as the cell under test. Level 1 relies on the use of covariance smoothing and Model Order Selection (MOS) rules to make inference on the number of jamming signals affecting each range bin within the radar range swath. Level 2 allows to discriminate among possible different interfering scenarios characterized by the same number of jammers via an unsupervised learning clustering fed by a suitable feature set. At the analysis stage, the performance of the devised architecture is investigated over simulated data to highlight the benefits of the approach.
Frequency Modulation Parasitic DOA Error Compensation
Sungdo Choi (Samsung Advanced Institute of Technology, Korea (South)); Hyunwoong Cho and Woosuk Kim (Samsung Advanced Institute of Technology (SAIT), Korea (South)); Minsung Eo and Seungtae Khang (Samsung Advanced Institute of Technology, Korea (South)); Jongseok Kim (Samsung Advanced Institute of Technology (SAIT), Korea (South))
This paper introduces a phase normalize transform (PNT) method to compensate for parasitic direction-of-arrival (DOA) error due to frequency modulation for wideband DOA estimation. Because the existing wideband DOA methods are based on a covariance alignment scheme, there is a limit to the location or order of DOA processing in the radar signal processing (RSP) framework. Because the proposed method compensates for parasitic components caused by the wideband signal, it is advantageous to apply it to various RSP frameworks without any modification. The performances of the popular wideband DOA method and narrowband MUSIC method are compared with that of the proposed method through computer simulations. The simulations demonstrate the higher performance of the proposed method in the mid-to-high SNR range compared with that of previous methods. In addition, real data obtained from a developed 79 GHz radar system are presented to show the effectiveness of the proposed method.
Mainlobe Deceptive Jammer Suppression with MIMO Radar Using Element-Pulse Coding
Lan Lan, Guisheng Liao and Jingwei Xu (Xidian University, China); Yuhong Zhang (Xidian University, USA)
This paper deals with the suppression of mainlobe deceptive jammers in a collocated multiple-input multiple-output (MIMO) radar by element-pulse coding (EPC). At the design stage, by simultaneously coding both spatial elements and slow time pulses, the EPC scheme has the capability of discriminating the true target from the false ones, corresponding to different range ambiguity regions. In such a way, by designing an appropriate coding coefficient, the false targets, from unwanted range ambiguous regions, can be suppressed by nulling at the equivalent transmit beampattern of the true target. At the analysis stage, a detection performance comparison is implemented among the proposed scheme and traditional radar frameworks. Numerical simulations are carried out to corroborate the theoretical developments.

ThA-P6: High resolution SAR/ISAR

Chair: Marco Martorella (University of Pisa, Italy)
A Computationally Efficient Approach for Velocity Estimation of Ground Moving Targets
Amir Hossein Oveis (Radar and Surveillance Systems (RaSS · Lab - CNIT), Italy); Marco Martorella (University of Pisa, Italy); MohammadAli Sebt (K. N. Toosi University of Technology, Iran); Ali Noroozi (K. N. Toosi University of Technology, Australia)
In this paper, a novel algorithm for the parameters estimation of chirp signals is proposed. Chirp rate and centroid frequency of chirp signals are estimated based on a one-dimensional dechirp optimization problem (DOP). The proposed DOP algorithm is useful for synthetic aperture radar (SAR) systems since the azimuth signal of a moving target represents chirp properties, so the along-track and across-track velocities of moving targets can be efficiently estimated. The distinctive feature of the proposed DOP algorithm, as compared with other motion parameters estimators, is its efficiency from the computational point of view. This is obtained by converting the traditional two-dimensional search to an efficient and simple one-dimensional optimization problem. Finally, simulations are presented to validate the theoretical investigations.
A Fast ISAR Imaging Method for Rapidly Spinning Targets Using Pseudo-Polar Coordinate in Range-Doppler Domain
Mengyi Qin and Dong Li (Chongqing University, China); Xi Luo (Xi’an Institute of Space Electronics and Information Technology, China); Hongqing Liu (Chongqing University of Posts and Telecommunications, China); Jinzhi Ren and Qihang Cao (Chongqing University, China)
Designing an efficient imaging approach is one key unsolved issue faced by inverse synthetic aperture radar (ISAR) imaging for spinning targets. In this paper, a fast ISAR imaging method based using pseudo-polar coordinate system in Range-Doppler (RD) domain is developed to solve this issue. First, a precise analytic expression is derived in RD domain utilizing the principle of stationary phase (POSP). Based on that, with a novel interpolation kernel function, the well-focused ISAR image for spinning target is sampled using a pseudo-polar grid. After that, the interpolation-free Stolt mapping algorithm is adopted to further reduce the computational complexity. Compared with the available imaging methods, the proposed method has advantages in computational cost and low signal-to-noise ratios (SNR) environment due to avoiding the multi-dimension searching operation and utilizing the two-dimensional (2-D) coherent integrated gain. Finally, the simulations are provided to verify the correctness of the proposed algorithm.
A Registration Strategy for Cicular SAR Noncoherent Imaging
Yuxiao Luo (National University of Defense Technology, China); Daoxiang An, Xiaotao Huang, Leping Chen and Wu Wang (National University of Defense Technology, China)
Circular synthetic aperture radar (CSAR) enables 360-degree observation of the scene. Compared with coherent imaging, noncoherent imaging can suppress speckle and obtain a large number of texture features. However, due to the existence of a man-made anisotropic targets and the unique imaging geometry of CSAR, the different subaperture images are quite different, which makes the registration very difficult. Therefore, this manuscript proposed a factorized registration strategy. First, the full aperture is divided into multiple subapertures; then the adjacent subaperture images are registered and the noncoherent superimposition of the subaperture images yields the next stage image; finally, the previous step is performed iteratively until the final imaging result is obtained. Experimental results based on Ka-band CSAR data verify the effectiveness of the strategy.
Three Dimensional Electromagnetic Vortex Radar Imaging Based on the Modified RD Algorithm
Lin Wang (University of Electronic Science and Technology of China, China); Lujing Tao (The 28th Research Institute China Electronic Technology Group Corporation, China); Zhongyu Li (University of Electronic Science and Technology of China, China); Junjie Wu (University of Electronic Science and Technology of China (UESTC), China); Jianyu Yang (School of Electronic Engineering, China)
The vortex electromagnetic (EM) wave carrying orbital angular momentum (OAM) is promising to benefit the higher degree of freedom for synthetic aperture radar (SAR) imaging. Combining the traditional two-dimensional (2D) SAR and the vortex EM wave, vortex SAR can obtain the three-dimensional (3D) information to realize 3D imaging. This letter proposes a novel 3D vortex SAR imaging method based on the vortex EM wave to obtain 3D target information. First, a vortex SAR imaging model is established by the proposed 3D imaging algorithm. Then, the modified range-doppler (RD) algorithm is therefore proposed to reconstruct the target's profiles. The objective of the 3D RD algorithm is to decompose the time-domain correlation operations of the SAR 3D space into three independent one-dimensional correlations and then perform separate focusing processing on each dimension. Finally, the 3D information of the targets can be obtained according to the modified RD algorithm. The point target simulation is performed in this paper. The simulation results establish the effectiveness of the proposed technique. This paper lays the groundwork for the development of vortex EM waves in the field of 3D radar imaging.
Sparse Reconstruction for Synthetic Aperture Radar Based on Split SPICE
Jiawei Luo, Yongchao Zhang, Deqing Mao, Yin Zhang and Yulin Huang (University of Electronic Science and Technology of China, China); Jianyu Yang (School of Electronic Engineering, China)
Synthetic aperture radar (SAR) technology is widely applied to observing the ground surface. Recently, the sparse iterative covariance-based estimator (SPICE) method has been developed for SAR to obtain a higher resolution and lower sidelobes of the sparse targets over conventional fast Fourier transform (FFT)-based method. However, the matrix inversion operation is required for each iteration, which results notably high computational complexity. Aiming at this problem, a split SPICE method is proposed to improve the computational efficiency. The algorithm is derived from the equivalence between the SPICE and the weighted square root least absolute shrinkage and selection operator (LASSO) problem. Furthermore, the problem is split into three sub-problems by introducing an auxiliary variable. The proposed split SPICE method has the advantages of higher computing efficiency over conventional SPICE algorithm. Compared with conventional Split Bregman method, the proposed method requires less user parameter. Simulations verify the excellent performance of the proposed method.
Road Surface Quality Assessment Using Polarimetric Airborne SAR
Arun Babu and Stefan V. Baumgartner (German Aerospace Center (DLR), Germany)
Since the road infrastructure is an important factor for the economy and safety, the continuous monitoring of the road condition is a necessity. Compared to the widely used costly, time consuming and labour-intensive road condition monitoring using measurement vehicles all over the country, the potential of SAR polarimetry to remotely monitor the road surface roughness, cracks and potholes are investigated in this study. The polarimetric analysis of fully polarimetric X-band radar datasets acquired over the Kaufbeuren test site with DLR's airborne sensor F-SAR revealed that the anisotropy and coherency matrix (T3) elements are sensitive to the road surface roughness and can be used to retrieve the vertical surface roughness. The cross-polar sigma nought images show a considerable increase in their magnitude over the cracks and potholes on the road surface. The initial experimental results obtained from this study are discussed in this paper.
Optimized Minimum-Search for SAR Backprojection Autofocus on GPUs Using CUDA
Niklas Rother, Christian Fahnemann and Jan Wittler (Leibniz Universität Hannover, Germany); Holger Blume (Leibniz Universitaet Hannover, Germany)
Autofocus techniques for synthetic aperture radar (SAR) can improve the image quality substantially. Their high computational complexity imposes a challenge when employing them in runtime-critical implementations. This paper presents an autofocus implementation for stripmap SAR specially optimized for parallel architectures like GPUs. Thorough evaluation using real SAR data shows that the tunable parameters of the algorithm allow to counterbalance runtime and achieved image quality.
Hybrid Conditional Random Fields Based on Complex-valued 3D CNN for PolSAR Image Classification
Peng Zhang, Beibei Li, Xiaofeng Tan and Yinyin Jiang (Xidian University, China); Ming Li (XIDIAN University, China); Yan Wu (Xidian University, China)
In this paper, we propose a hybrid conditional random fields based on complex-valued 3-dimensional convolutional neural network (CV3DCNN), named as HCRF-CV3DCNN, for polarimetric synthetic aperture radar (PolSAR) image classification. HCRF-CV3DCNN extracts deep features of PolSAR image by CV3DCNN, and thus exploits amplitude and phase information of PolSAR data to measure the class probability in the framework of random fields. Additionally, based on the CV3DCNN-based class probabilities, the relative entropy of class distributions is derived and utilized to regulate the label interactions to enhance the accuracy of edge location in the classification. At last, to capture PolSAR image information in a more complete manner, the deep features and PolSAR scattering statistics are integrated into HCRF-CV3DCNN based on Bayesian fusion. In this way, HCRF-CV3DCNN effectively combines the representation-learning ability of deep learning model with the modeling power of random fields including spatial correlation and data statistics. The experimental results demonstrate the superiority of HCRF-CV3DCNN over the recent deep learning models for PolSAR image classification.
Improved Frequency Division Algorithm for Residual Motion Errors in Synthetic Aperture Radar
Qianrong Lu (Shanghai Radio Equipment Research Institute, China)
In high-resolution synthetic aperture radar, the impacts of complicated motion errors on azimuth phase should be eliminated to obtain well-focused images. To this end, an improved frequency-division has been developed based on the precise angle-frequency relation. The proposed method takes into account both the cross-track and along-track motion errors and could completely compensate the residual aperture-variant motion. Moreover, this fashion could be integrated into traditional image formation algorithms with no changes. After imaging, a paired-echo suppression method based on phase gradient autofocus with integrated notch filter is presented. Both the simulated data and real airborne data have been applied to validate this method.
Vector Extrapolation Based Fast Split Bregman Algorithm for Radar Forward-Looking Imaging
Qiping Zhang, Yin Zhang, Yongchao Zhang and Yulin Huang (University of Electronic Science and Technology of China, China); Jianyu Yang (School of Electronic Engineering, China); Qingying Yi (University of Electronic Science and Technology of China, China)
Conventional split Bregman algorithm (SBA) has been used for improving the azimuth resolution for radar forward-looking imaging. However its convergence speed is not satisfactory. In this paper, a vector extrapolation based fast SBA (VEFSBA) is presented to accelerate the convergence of traditional SBA and improve azimuth resolution of radar forward-looking imaging. Based on the principle of vector extrapolation, the proposed VEFSBA adopts the second-order vector extrapolation strategy to reduce the number of iterations of traditional SBA. It not only effectively improves the azimuth resolution of radar forward-looking imaging, but also greatly reduces the number of iterations compared with the traditional SBA. Simulation and measured data are presented to demonstrate the superior performance of the proposed VEFSBA.
A Modified CSA for Missile-borne SAR with Curved Trajectory
Zhang Yun and Lu Chenyue (Harbin Institute of Technology, China); Haojian Zhang (Beijing Institute of Aerospace Systems Engineering, China); Hongbo Li (Harbin Institute of Technology, China)
The missile's flight trajectory in the final guidance section is complex and generally has a high diving velocity. Therefore, the missile-borne SAR is faced with the problems of high maneuverability, high squint mode, and curved trajectory, conventional SAR imaging algorithms are no longer applicable. In this paper, a modified CSA suitable for wide scenes is proposed. Firstly, using the method of series reversion derive the two-dimensional spectrum, and then analyzing the range bending factor under wide scene condition which determine the scaling equation with curve trajectory. Based on this, a modified CSA algorithm can be derived. The simulation results show that this algorithm can achieve high-precision imaging under the condition of diving maneuver.

ThA-P7: Detection and estimation

Chairs: Bernard Mulgrew (Institute for Digital Communications, The University of Edinburgh, United Kingdom (Great Britain)), Piotr Samczynski (Warsaw University of Technology, Poland)
A Coarse-to-Fine Robust Estimation of FMCW Radar Signal for Vital Sign Detection
Kang Liu and Yuanhui Zhang (China Jiliang University, China)
This paper presents a robust vital sign detection system, which uses 77GHz mmWave radar sensor. Two essential signals breathing and heart beat rate are estimated by measuring the change in phase of FMCW signal with time. Breathing harmonics have significant impact on the heart beat spectrum. In order to extract heart beat signal, a coarse-to-fine estimation approach is developed. The first contribution of this approach is to predicts the heart beat rate in the time domain of the unwrapped phase signal, where an adaptive threshold is calculated to filter invalid wave peak-valley pairs. Furthermore, observation window size selection for the vital sign detection is optimized by two criteria: peak difference and local SNR. The second contribution is to estimate heart beat rate in spectral domain based on the predicted rate. Experiments show that this coarse-to-fine approach gives a robust estimate of heart rate as well as breathing rate.
A Parameter-Controlled Rx Beamspace Transformation Design Method
Nadav Neuberger and Risto Vehmas (Fraunhofer FHR, Germany); Joachim H. G. Ender (Fraunhofer FHR & Universität Siegen, Germany)
Sensitive target detection and accurate direction-of-arrival (DOA) estimation are key goals of the modern phased array radar designer. To this end, a high number of receiving antenna elements is necessary. This inevitably leads to a high computational complexity for the required signal processing schemes. Therefore, common solutions reduce the data dimension by transforming the element level data into a lower dimensional beamspace. However, conventional transformation methods are sub-optimal and suffer from inefficient use of resources. Moreover, these methods usually focus on detection performance rather than DOA estimation accuracy. This paper presents an analytical formulation for a beamspace transformation jointly achieving optimal detection and DOA estimation performance within a spatial sector of interest. We present a novel design method enabling a trade-off between detection and DOA estimation performance when resources are insufficient. Comparative studies concerning theoretical and empirical results illustrate the superiority and practical aspects of the proposed design method.
Coherent Detection for Weak Target Based on Frequency Phase Compensation Function
Lanjin Lin (University of Electronic Science and Technology of China, China); Guohao Sun (Sichuan University, China); Ziyang Cheng (University of Electronic Science and Technology of China, China); Zi-Shu He (University of Electronics Science and Technology of China, China)
This paper addresses the long-time coherent integration detection problem for a weak maneuvering target, involving linear range migration (LRM), range curvature (RC), and Doppler frequency migration (DFM). To solve the above problems, an efficient coherent integration method based on frequency phase compensation function process (FPCFP) is proposed. By jointly searching the unknown motion parameters in velocity and acceleration domains, the FPCFP algorithm not only can eliminate LRM, RC, and DFM simultaneously, but also can realize the coherent integration. Compared with the existing algorithms, the proposed method achieves a better detection probability in a low signal-to-noise ratio (SNR) environment and ensures high computational efficiency. Several experiments are performed to verify the effectiveness of the presented approach.
Doppler Signature Separation of Mixed O/X-Mode over-the-Horizon Radar Signals
Ammar Ahmed and Yimin D. Zhang (Temple University, USA); Braham Himed (AFRL, USA)
In this paper, we consider the Doppler signatures of local multipath target signals in an over-the-horizon radar and the objective is to separate signals corresponding to the ordinary (O) and extraordinary (X) propagation modes. As signals of these two modes are reflected at different ionosphere heights, the rendered Doppler signatures are more complicated and cannot be directly analyzed for parameter estimation using conventional local multipath models which are developed for a single propagation mode. We consider a mixed O/X-mode signal model and analyze the resulting Doppler signatures comprising of more signal components. It is shown that sparsity-based methods estimate the Doppler components with improved resolution and accuracy. Moreover, the proposed group sparsity-based strategy enables separation of the resolved Doppler components corresponding to the two propagation modes. We present the simulation results for a challenging scenario where the multipath Doppler signatures corresponding to the two modes are interlaced.
Moving Target Detection for Asynchronous Data with Distributed MIMO Radar
Shixing Yang, Wei Yi, Yangming Lai and Yao Wang (University of Electronic Science and Technology of China, China)
In this paper, we consider the problem of moving target detection with multiple-input multiple-output (MIMO) radar under the "focused-transmit focused-receive" operating model. Since one transmitter may illuminate the moving target at a different moment with others, the received data of all receivers for those transmitters are asynchronous. To solve this problem, we propose a method named states exhausting detection (SED) that detects the moving target with asynchronous data provided by multiple transmitters. For SED, we first set several possible velocities of one target. Then we calculate the target states illuminated by all transmitters for every velocity. Finally, the presence of each target can be determined by exhausting all target states. Both analysis and extensive simulation results are provided to corroborate the proposed radar system and algorithm.
An Efficient Multi-PRF and Multi-frame Joint Detection Approach in Multi-PRF Radar System
Wujun Li, Ph Zhang, Lei Chai, Kezhu Liu, Qiyun Peng and Wei Yi (University of Electronic Science and Technology of China, China)
This paper mainly addresses the weak target detection and tracking problems in multiple pulse repetition frequency (MPRF) radar system using the multi-frame track-before-detect (MF-TBD) method. Generally, each measurement frame received by the MPRF radar faces the range and Doppler ambiguities at the specific pulse repetition frequency (PRF) setting. To achieve this goal, we first propose a multi-PRF and multi-frame TBD (MPMF-TBD) algorithm, which provides an energy-efficient solution to accomplish the target kinematic state estimation and multi-PRF ambiguity resolution simultaneously. Then, by employing cross-boundary multi-frame integration and parallel processing structure, the multi-PRF and multi-frame integration process is improved, thereby further reducing the computational complexity of MPMF-TBD algorithm. Simulation results demonstrated the efficacy of the proposed algorithm.
Signal Fusion-Based Detection with an Intuitive Weighting Method
Jing Lu (Xidian University & National Laboratory of Radar Signal Processing, China); Shenghua Zhou (National Laboratory of Radar Signal Processing, Xidian University, China); Jianlai Wang, Dawei Li and Tieshan Feng (China Academy of Launch Vehicle Technology, China); Hongwei Liu (National Laboratory of Radar Signal Processing, China); Hui Ma (Xidian University, China)
In a distributed multiple-input multiple-output (MIMO) radar, signal fusion based detection can achieve the best detection performance but the global test may be a complicated combination of local observations, such that a closed-form expression of the false alarm rate may be hard to develop, especially for heterogeneous local observations. In this paper, we study how to impose better linear weights on local test statistics for a global test resulting in a better detection performance. Since the weights of the generalized likelihood ratio test (GLRT) have been proved to be suboptimal, an intuitive and tractable weighting method is presented that weights the local test statistics according to the estimated channel signal-to-noise ratios (SNRs) and the power of the local test statistics measured by the detection probability curve. With a linear combination of local test statistics, the false alarm rates can be computed conveniently. This method is also applicable to heterogeneous local observations. Numerical results for the performance of the distributed cell-averaging constant false alarm rate (CA-CFAR) algorithm and the distributed GLRT algorithm indicate that the proposed weighting method can outperform conventional algorithms.
Distributed Radar Multi-Frame Detection with Least Squares Quantization
Jing Lu (Xidian University & National Laboratory of Radar Signal Processing, China); Shenghua Zhou (National Laboratory of Radar Signal Processing, Xidian University, China); Pramod Varshney (Syracuse University, USA); Jibin Zheng (Xidian University, China); Xiaojun Peng (China Academy of Launch Vehicle Technology, China); Hongwei Liu (National Laboratory of Radar Signal Processing, China); Zhiqiang Shao (Xidian University, China)
In a distributed multiple-input multiple-output (MIMO) radar system, multiple-frame local observations can be transmitted to a fusion center (FC) for a better detection performance, but the communication cost may be huge. In this paper, we study how to impose the least squares quantization (LSQ) method on distributed multi-frame detection (MFD). Local test statistics instead of raw signals are quantized by the LSQ algorithm and then a decision rule is formulated based on the dynamic-programming based MFD algorithm. Numerical results indicate that the LSQ algorithm causes an insignificant detection performance loss at three-bit LSQ quantization. Meanwhile, this method greatly reduces the computational complexity and the communications bandwidth costs.
Incoherent Change Detection Methods for Wavelength-Resolution SAR Image Stacks Based on Masking Techniques
Dimas Irion Alves and Crístian Müller (Federal University of Pampa, Brazil); Vu Viet Thuy and Mats I. Pettersson (Blekinge Institute of Technology, Sweden); Pablo Kunz de Jesus (Instituto Tecnologico de Aeronautica - ITA, Brazil); Renato Machado (Aeronautics Institute of Technology, Brazil); Bartolomeu F. Uchôa-Filho (Federal University of Santa Catarina & Communications Research Group, Brazil)
This paper presents two incoherent change detection methods for wavelength-resolution synthetic aperture radars (SAR) image stacks based on masking techniques. The first technique proposed is the Simple Masking Detection (SMD). This method uses the statistical behavior of pixels-sets in the image stack to create a binary mask, which is used to remove pixels that are not related to changes in a surveillance image from the same interest region. The second technique is the Multiple Concatenated Masking Detection (MCMD), which produces a more selective mask than the SMD by concatenating multiple masks from different image stacks. The MCMD can be used in specific applications where multiple stacks share common patterns of target deployments. Both proposed techniques were evaluated using 24 incoherent SAR images obtained by the CARABAS II system. The experimental results revealed that the proposed detection methods have better performance in terms of probability of detection and false alarm rate when compared with other change detection techniques, especially for high detection probabilities scenarios.
A Genetic Approach to MTI Filter Design with Nonuniform Sampling
Abbas Ghorbani, Seyed Mohammad Karbasi and Mohammad Mahdi Nayebi (Sharif University of Technology, Iran)
In modern radar systems, moving target indicator (MTI) filters are generally used to remove the echo caused by fixed unwanted targets, as well as signal-dependent interference (i.e., clutter). The MTI filters are simply defined as the weighted sum of pulses with tuned periods. Namely, the design parameters that mainly characterize such filters are the weighting coefficients and interpulse periods (IPPs). In this paper, a two-step algorithm is introduced to calculate the design parameters. As to the weighting coefficients, the deviation of the designed passband response is optimized using the least square criterion. As to the IPPs, the minimum of the signal-to-clutter ratio (SCR) in the passband is maximized utilizing the genetic algorithm. In order to make control over the stopband, particular constraints is enforced to induce nulls at limited numbers of frequency spots in the clutter rejection band. The performance assessments illustrates the effectiveness of the genetic least square (GLS) method, as compared with other existing ones.
Structure-Based Sensing Matrix Optimization for Extended Target Ranging
Yating Wang, Feng Xi and Shengyao Chen (Nanjing University of Science and Technology, China)
In compressive sensing radar applications, sensing matrix optimization is important to improve the reconstruction performance. In this paper, prior knowledge of the signals of interest is exploited to design sensing matrix for extended targets. We separate the coherence between different columns of the equivalent matrix into two different sets based on the structure information of extended targets. Our method is to shrink the coherence corresponding to the different sets under different pre-set thresholds. To improve the robustness of our problem, a Frobenius norm regularization term is also added. An alternative algorithm based on gradient descent method is devised to solve the optimization problem. Simulation results demonstrate that the proposed sensing matrix significantly improves the reconstruction performance by exploiting target structure information.
Derivation and Performance of an MLE Based Path Loss Augmented Correlation Radar Receiver
Jase Furgerson and Dinesh Rajan (Southern Methodist University, USA)
In this paper, we derive Path loss Augmented Correlation (PAC), a Maximum Likelihood Estimator (MLE) for the target range in a radar system that also incorporates the path loss, as a function of propagation delay. The PAC receiver is tested for Monostatic, Multistatic, and MIMO systems and is compared to traditional correlation receivers that use only the propagation delay to estimate the range. These experiments reveal substantial reduction in range error (24-49%) for a system that has been calibrated to the path loss exponent. The robustness of the PAC receiver to the knowledge of the path loss exponent is also quantified.