Program for 2020 IEEE Sensors Applications Symposium (SAS)

Monday, March 9 15:50 - 17:30

MA1: Medical and Biomedical Applications

Room A
Chair: Seohyun Lee
15:50 Metrology in eye pressure measurements
Measurements and their precision are crucial for many fields of industry and services that have a direct impact on the quality, safety and reliability of products and applications. One of the areas that have a particular effect on national finances and people is the medical sector. The reliability and precision of the measured quantities have a direct and significant effect on the final diagnosis of a patient. An unrepresentative measurement of subject's true essential indicators like blood pressure, body temperature or heart rate could lead to false diagnosis and incorrect or unnecessary treatment. This can in most serious cases result in a decrease of a patients quality of life and even in shortened live span. The treatment brings of course financial and social aspects that all affects the country's economy negatively. This paper focuses on the improvement of medical measurements by applying metrological procedures that enable a more precise and reliable measurements. The subject of these efforts is the measurement of eye pressure by non-contact means that are an initial diagnostic indicator for the development of eye glaucoma. The general problematic of the measurements together with the theoretical and practical solution by means of a standard is presented.
Presenter bio: Worked as research scientist in the field of metrology for more than 7 years in the national metrology laboratories of Slovakia (SMU).Has been and is involved in multiple international research projects with focus on metrology. Is the representative and member in technical committees and working group in EURAMET and BIPM.
16:10 Visualization and Data Analysis for Intracellular Transport using Computer Vision Techniques
Internalization of nanoparticles into intracellular area includes key information in biomedical field, such as cell signal pathway and drug delivery. Although the tracking of the individual nanoparticles in the cytoplasmic area has revealed the movement of the target in terms of single-particle level, the whole cell-level study is fundamental in order to efficiently acquire a large dataset of intracellular transport. In the present study, visualization and data analysis methods for understanding the entire cell-level intracellular transport in a living cell is suggested, by applying computer vision techniques to the cell images collected on the camera image sensor. Using the changes in the optical flow of the quantum dot-labeled vesicles for the entire intracellular area, our method showed the possibility of the time series analysis of vesicle movement related to the transport by two different types of molecular motors, dynein and kinesin.
Presenter bio: Seohyun Lee received her Ph.D in Physics from the University of Tokyo, Japan. Currently, she is a Project Researcher in the department of Information Physics and Computing at the University of Tokyo. Her research interest covers imaging microscopy, computer vision and image processing, and machine learning.
16:30 Preliminary investigation into low-cost stretch sensors for stomach deformation measurement
Physical simulation of gastric tract motility, through the use of controlled flexi-walled reactors, can benefit from measured feedback of the amount of stretch that regions of the deformable membrane wall undergo. Appropriate sensors need to be at least as flexible as the surface they are mounted on, and have the ability to stretch to the same extent. One method for measuring the stretch of highly flexible membrane walls is through the use of conductive ionic fluids encased in a flexible elastomer. A change in resistance can be measured between two terminals of the ionic liquid pathway as the sensor strip undergoes stretch deformation. A simple, low-cost approach to fabricating this form of sensor is to encase a conductive saline solution in a highly flexible silicone tube, with electrodes placed in contact with the fluid at the two ends of the sealed tube. Initial results indicate that this basic approach provides good stability and repeatability of resistance measurement readings during stretching. Further testing of the characteristics of this type of sensor is required to properly assess its capabilities against commercial stretch sensors and state-of-the-art devices.
Presenter bio: G. Sen Gupta graduated with a BE (Electronics) from the University of Indore, India in 1982, MEE from the University of Eindhoven, The Netherlands in 1984 and a PhD in Computer Systems Engineering in 2008 from Massey University, New Zealand. In 1984 he joined Philips India and worked as an Automation Engineer till 1989. Thereafter he worked as a Senior Lecturer in the School of Electrical and Electronic Engineering at Singapore Polytechnic, Singapore. He is with Massey University, New Zealand, since September 2002 where he is a Professor. He has published over 160 papers in international journals & conference proceedings, co-authored 2 books on programming, edited 4 conference proceedings, and edited 2 special issues of international journals (IEEE Sensors Journal and IJISTA) as guest editor. He is a Senior Member of IEEE. His area of interest is robotics, autonomous systems, embedded systems, sensor applications and vision processing for real-time applications.
16:50 Detection of Mathematical Fluency Effects on Working Memory using Near Infrared Spectroscopy
The prefrontal cortex (PFC) not only plays a less mathematics-specific role in number processing and calculation but it is also responsible for working memory demands coping. Through functional near-infrared spectroscopy imaging, this study explored how differently the PFC of 31 subjects with contrasting mathematical fluency would respond to increasing working memory demands of mental arithmetic task. Area under the curve analyses of the oxygenated hemoglobin signals, hereinafter representing brain activation, were compared across levels and groups. Depending on task performance, the subjects were allocated to either normal performers (NP) or high performers (HP) group. Both groups showed sensitivity with regards to task performance (i.e., the number of problems attempted and accuracy) towards increasing working memory demands (level 1 to 3). With increasing task level, NP's PFC activation increased while HP's remained unchanged. NP also showed consistent greater PFC activation than HP in all three levels. These findings implied that the PFC plays a more auxiliary role to cope with working memory demands imposed by mental arithmetic on individuals who are less fluent in mathematics. This can potentially serve as a framework to monitor workload hemodynamics continuously and unobtrusively.
Presenter bio: Wei Chun Ung is a student at the Universiti Teknologi PETRONAS. He is currently pursuing his PhD in Electrical & Electronic Engineering by Research. He completed his seven-month internship in optical topography at Hitachi Ltd., Research & Development Group, Center for Exploratory Research, Japan. His current research interests include neurofeedback, and clinical applications of functional near-infrared spectroscopy, particularly involving dementia.
17:10 A PPG-ECG Combo System for the Monitoring of the Aging State of Arteries
Arterial hypertension is an indicator of the cardiovascular pathologies. It causes arterial stiffness which worsen with the aging. This paper presents a PPG-ECG combo measurement system, using a novel signal processing methodology, which aims at the evaluation of the Pulse Wave Velocity in order to obtain clinical parameters of the aging state of the arteries (Arterial Stiffness). Preliminary experimental results are presented, which demonstrates the suitability of the proposed methodology.

MB1: Internet of Things

Room B
Chairs: Alessandro Depari, Divya Lohani
15:50 CoAP + DTLS: A Comprehensive Overview of Cryptographic Performance on an IOT Scenario
Internet of things (IoT) and Fog computing applications deal with sensitive data and need security tools to be protected against attackers. CoAP (Constrained Application Protocol) combined with DTLS (Datagram Transport Layer Security) provide security to IoT/Fog applications, however processing times need to be considered when using this combination due to IoT/Fog environment constraints. Our work presents a CoAP with DTLS application and analyzes the performance of Raspberry Pi 3 during DTLS handshakes, data encryption and data decryption with the most relevant cipher suites. The performance of confirmable and non-confirmable CoAP POST requests is also measured and discussed in our work. We discovered that cipher suites that use RSA as an authentication method on handshake are slightly faster than cipher suites that use ECDSA, while symmetric key encryption with AES256(128)-GCM are 40% faster than AES256(128) default modes. Our study also suggests CoAP modifications to obtain higher efficiency and it might help future IoT/Fog application developers to understand CoAP and DTLS union, providing an application example and performance metrics.
Presenter bio: Bachelor in Computer Science since 2019. Master's student at the postgraduate program in Computer Science from the Federal University Of Santa Catarina, in Brazil. Researches in the IoT Security field.
16:10 A self-organizing efficient power generation system in extreme condition for Waggle
To deploy a novel wireless sensor platform in non-urban area, great consideration must be taken on its powering system. This paper describes an autonomous power generation system in extreme weather such as winter cold, summer heat, and other natural events. The system was applied to Waggle. Performance of the system was evaluated by measuring the battery voltage. The Waggle Platform, a research project from Argonne National Laboratory is one of the most practical autonomous nodes allowing the collection of environmental data in the urban area. Previous work only achieved with the limitation of 3.17 hours of operation time when all modules were on. Our power system lasted for 8 hours which is 2.5 times more than the previous one. This paper examines the power system to provide continuous power and remain operational 24 x 7 x 365, regardless of the environmental conditions. Through this project, we tested the improved structure of the power system.
Presenter bio: Soongsil University student.
16:30 IoT Enabled Low Cost Air Quality Sensor
Air pollution poses significant risks to environment and health. Air quality monitoring stations are often confined to a small number of locations due to the high cost of the monitoring equipment. They provide a low fidelity picture of the air quality in the city; local variations are overlooked. However, recent developments in low cost sensor technology and wireless communication systems like Internet of Things (IoT) provide an opportunity to use arrayed sensor networks to measure air quality, in real time, at a large number of locations. This paper reports the development of a novel low cost sensor node that utilizes cost-effective electrochemical sensors to measure Carbon Monoxide (CO) and Nitrogen Dioxide (NO2) concentrations and an infrared sensor to measure Particulate Matter (PM) levels. The node can be powered by either solar-recharged battery or mains supply. It is capable of long-range, low power communication over public or private LoRaWAN IoT network and short-range high data rate communication over Wi-Fi. The developed sensor nodes were co-located with an accurate reference CO sensor for field calibration. The low-cost sensors' data shows strong correlation with the data collected from the reference sensor. Offset and gain calibration further improves the quality of the sensor data.
Presenter bio: Baden is a PhD student studying at Massey University, New Zealand. His doctoral research focuses on automating grape yield estimation in vineyards using 3D camera technology. In addition to these studies, Baden is also involved in research on indoor localization using visible light, acoustic imaging using active and passive arrays, dense urban IoT sensor networks for air quality monitoring, and robot design and locomotion strategies.
16:50 Modeling IoT Enabled Automotive System for Accident Detection and Classification
Millions of people get injured, disabled or die in automotive accidents each year. Knowledge about the type of road accident is invaluable to the emergency medical services providers for optimal planning and execution of the rescue operation. An IoT based system has been developed in this work to report the occurrence, location as well as the type of road accident. The system uses in-built sensors of passenger smartphone to detect and classify the accident as head-on collision, rollover or fall-off. The accuracy of the proposed system, which uses Naïve Bayes classifier for classification, has been evaluated using precision, recall, F1 score and ROC curve.
Presenter bio: Dr. Divya Lohani is currently an Assistant Professor in the Department of Computer Science & Engineering at Shiv Nadar University, India. She is a researcher in the field of Internet of Things, Wireless Sensor Networks and Sensor Data Analytics.
17:10 RTK-LoRA: High-Precision, Long-Range And Energy-Efficiency Localization for Mobile and Self-sustainable IoT devices
High precision Global Navigation Satellite System (GNSS) is a crucial feature for geo-localization to enhance future applications such as self-driving vehicles. Real-Time Kinematic (RTK) is a promising technology to achieve centimeter precision in GNSS. However, it requires radio communication, which usually is power-hungry and costly, e.g. when using the 4G network. Hence, today RTK is not much exploited in low power energy-efficient devices. In this work, we present a novel sub-meter precision RTK-base system that also achieves the requirements of low power and energy efficiency. The proposed system exploits a novel GNSS module with RTK combined with a long-range and low-power radio (LoRa) to achieve geolocalization with minimal wireless radio infrastructure requirements. We evaluate three different GNSS modules and compare their performance in terms of power and especially precision. Experimental results, with in-field measurements, show an average accuracy of tens of centimeters with a single base station as geostationary reference anchor placed at kilometers of distance from the end-node performing the distance measurement. The peak accuracy measured was below 10cm.
Presenter bio: Philipp Mayer (S'17-M'18) received his B.Sc. degree in electrical engineering and information technology from the Technical University Vienna, Vienna, Austria in 2016. He is currently working toward a consecutive M.Sc. degree at the ETHZ, Zurich, Switzerland that focuses on energy neutral sensing for IoT. His research interests include low-power system design, unconventional methods for energy harvesting, wireless optogenetic and edge computing.
Mayer Philipp

Tuesday, March 10 9:00 - 10:20

TA1: Smart Sensing: Machine Learning and Embedded Processing

Special Session
Room A
Chairs: Viviana Crescitelli, Michele Magno
9:00 Euclidean Distance based Loss Function for Eye-Gaze Estimation
The Loss function is an integral component in a Neural network. It affects the performance of CNN network in its classification. In this paper, we propose a Euclidean distance based Loss function for the CNN model, in an eye-gaze memory card game. We compared the Euclidean distance loss function with the well-known cross-entropy loss function. The performance parameters used in our comparison are prediction accuracy and average Euclidean distance prediction error. The results show that cross-entropy has better prediction accuracy. However, the Euclidean distance loss function provides a better average Euclidean distance prediction error resulting in better user experience. This is because the wrongly predicted eye gaze cards are near to the user intended card. In the case of cross-entropy, the predicted card error is quite evenly spread across the screen.
Presenter bio: Bu Sung Lee received his B.Sc. (Hons) and PhD from the Electrical and Electronics Department, Loughborough University of Technology, UK in 1982 and 1987 respectively. He is currently an Associate Professor with the School of Computer Engineering, Nanyang technological University. In addition, to the academic activities he also holds a Joint appointment as Director, Service Platform Lab, HP Labs. Singapore from July 2010 – June 2012. He has been actively involved with the Asia-Pacific research and education network since the formation of Singapore Advance Research and Education Network(SingAREN). He is the founding president of SingAREN society, 2003-2007, and is currently the President of SingAREN(2011-2015). Since 2004, he is a member of the technical management team of Trans-Eurasia Information Network (TEIN-2); the first large scale is the first large-scale research and education network for the Asia-Pacific. It connects eighteen countries in the region, and provides direct connectivity to Europe’s GÉANT2 network. Bu-Sung Lee has published over 300 peer preview conference papers, and 100 journal papers. His research areas cover both Grid/Cloud Computing and network. His particular interest are in data replication, scheduling, network Qos (wired and wireless), and ad hoc network. He has received a number of research grants. He is the co-author of a number of papers that has won Best Paper awards. He has given numerous Invited/Keynote address in conferences and played an active role in the academic community in organizing conferences.
9:20 An RGB/Infra-Red camera fusion approach for Multi-Person Pose Estimation in low light environments
Enabling collaborative robots to predict the human pose is a challenging, but important issue to address. Most of the development of human pose estimation (HPE) adopt RGB images as input to estimate anatomical keypoints with Deep Convolutional Neural Networks (DNNs). However, those approaches neglect the challenge of detecting features reliably during night-time or in difficult lighting conditions, leading to safety issues. In response to this limitation, we present in this paper an RGB/Infra-Red camera fusion approach, based on the open-source library OpenPose, and we show how the fusion of keypoints extracted from different images can be used to improve the human pose estimation performance in sparse light environments. Specifically, OpenPose is used to extract body joints from RGB and Infra-Red images and the contribution of each frame is combined by a fusion step. We investigate the potential of a fusion framework based on Deep Neural Networks and we compare it to a linear weighted average method. The proposed approach shows promising performances, with the best result outperforming conventional methods by a factor 1.8x on a custom data set of Infra-Red and RGB images captured in poor light conditions, where it is hard to recognize people even by human inspection.
Presenter bio: Engineer with broad international experience. Currently working as a scientific researcher in Tokyo at Hitachi Research and Development. Her research interest include the implementation of DNN techniques suitable for real time performance on edge devices.
Viviana Crescitelli
9:40 Multi-modality sensor fusion for gait classification using deep learning
Human gait has been acquired and studied through modalities such as video cameras, inertial sensors and floor sensors etc. Due to many environmental constraints such as illumination, noise, drifts over extended periods or restricted environment, the classification f-score of gait classifications is highly dependent on the usage scenario. This is addressed in this work by proposing sensor fusion of data obtained from 1) ambulatory inertial sensors (AIS) and 2) plastic optical fiber-based floor sensors (FS). Four gait activities are executed by 11 subjects on FS whilst wearing AIS. The proposed sensor fusion method achieves classification f-scores of 88% using artificial neural network (ANN) and 91% using convolutional neural network (CNN) by learning the best data representations from both modalities.
10:00 HR-SAR-NET: A Deep Neural Network for Urban Scene Segmentation from High-Resolution SAR Data
Synthetic aperture radar (SAR) data is becoming increasingly available to a wide range of users through commercial service providers with resolutions reaching 0.5 m/px. Segmenting SAR data still requires skilled personnel, limiting the potential for large-scale use. We show that it is possible to automatically and reliably perform urban scene segmentation from next-gen resolution SAR data (0.15 m/px) using deep neural networks (DNNs), achieving a pixel accuracy of 95.19% and a mean IoU of 74.67% with data collected over a region of merely 2.2km2. The presented DNN is not only effective, but is very small with only 63k parameters and computationally simple enough to achieve a throughput of around 500 Mpx/s using a single GPU. We further identify that additional SAR receiver antennas and data from multiple flights massively improve the segmentation accuracy. We describe a procedure for generating a high-quality segmentation ground truth from multiple inaccurate building and road annotations, which has been crucial to achieving these segmentation results.
Presenter bio: Xiaying Wang received her B.Sc. and M.Sc. degrees in biomedical engineering from Politecnico di Milano, Italy and ETH Zürich, Switzerland in 2016 and 2018, respectively. She is currently pursuing a Ph.D. degree at the Integrated Systems Laboratory at ETH Zürich. Her research interests include biosignal processing, low power embedded systems, energy-efficient smart sensors and machine learning on microcontrollers. She received the excellent paper award at the IEEE Healthcom conference in 2018 and she won the Ph.D. Fellowship funded by Swiss Data Science Center in 2019.

Tuesday, March 10 10:20 - 11:40

TPS: SAS 2020 Poster Session

Room: Foyer
Chairs: Alessandro Depari, Alain Pegatoquet
Adaptive Automatic Controller for Swing Assist by Pneumatic Artificial Muscle
Pneumatic Artificial Muscle (PAM) is essential to support patients with weak muscles in daily activities and physical rehabilitation process. Hence, the research on PAM as an assistive device is crucial for elderly and patients in the hospital, especially patients suffering from hemiplegic gait. Based on the understanding of the lower-limb muscle activation and human gait analysis, we proposed PAM Automatic Controller to drive PAM adaptively to support patient's gait in real-time. Automatic Controller primarily leverages the swing phase during gait to set the start of activation and muscle contraction duration to control the PAM using the smartphone wireless PAM driver. An evaluation experiment on Automatic Controller was conducted against the manual settings of PAM. Compared with manual setting, the Automatic Controller provides comparable to the best performance of static settings. In addition, it adjusts its contraction time in real-time without having the participants to carry out the calibration by running the 13 cases to set the parameters manually.
Presenter bio: Bu Sung Lee received his B.Sc. (Hons) and PhD from the Electrical and Electronics Department, Loughborough University of Technology, UK in 1982 and 1987 respectively. He is currently an Associate Professor with the School of Computer Engineering, Nanyang technological University. In addition, to the academic activities he also holds a Joint appointment as Director, Service Platform Lab, HP Labs. Singapore from July 2010 – June 2012. He has been actively involved with the Asia-Pacific research and education network since the formation of Singapore Advance Research and Education Network(SingAREN). He is the founding president of SingAREN society, 2003-2007, and is currently the President of SingAREN(2011-2015). Since 2004, he is a member of the technical management team of Trans-Eurasia Information Network (TEIN-2); the first large scale is the first large-scale research and education network for the Asia-Pacific. It connects eighteen countries in the region, and provides direct connectivity to Europe’s GÉANT2 network. Bu-Sung Lee has published over 300 peer preview conference papers, and 100 journal papers. His research areas cover both Grid/Cloud Computing and network. His particular interest are in data replication, scheduling, network Qos (wired and wireless), and ad hoc network. He has received a number of research grants. He is the co-author of a number of papers that has won Best Paper awards. He has given numerous Invited/Keynote address in conferences and played an active role in the academic community in organizing conferences.
Yield Process Control based on the Production Data
In semiconductor device manufacturing, there are many sophisticated processes involved in production. These processes need to be controlled precisely to meet the required specifications from customers. Therefore, it is critical to employ a process control system (PCS) to monitor and/or control variation in the manufacturing process. In this paper, we perform a statistical analysis of historical PCS data to: (1) determine which PCS parameters correlate with the product yield; and (2) perform regression analysis to estimate the yield from PCS data. As as result, we are able to identify key PCS parameters, namely time above liquidus, raising ramp rate, flux activation time and epoxy queue time, that can be optimized to improve the product yield.
A comparative study of LSTM and ARIMA for energy load prediction with enhanced data preprocessing
Energy load prediction plays a central role in the decision-making process of energy production and consumption for smart homes with systems based on energy harvesting. However, forecasting energy load turned out to be a difficult problem since time series data used for the prediction involve both linear and non-linear properties. In this paper, we proposed a system which can predict a daily future energy load in a smart home based on LSTM and ARIMA models. To improve the energy load forecasting accuracy, we propose a new data preprocessing algorithm called STDAN (Same Time a Day Ago or Next) to fill the missing values. This technique is compared with well-known techniques using previous or mean values. A comparison between LSTM and ARIMA is provided for short and medium-term load forecasting. Results show that LSTM outperforms ARIMA in all cases. Finally, we also evaluated our training model based on LSTM with a new data set and the model provides an around 80% accuracy.
Fabrication of a Pseudo-reference Electrode on a Flexible Substrate and Its Application to Heavy Metal Ion Detection
Flexible electrochemical sensors with three-electrode configuration are useful for various in-situ measurements of electroactive species. An electrochemical modification of a printed silver layer is presented as a method to integrate a mechanical and electrochemically stable pseudo-reference electrode. The surface morphologies and compositions of the pseudo-reference electrode were characterized by a scanning electron microscope, an energy dispersive X-ray spectroscopy (EDS) analysis, and an X-ray photoelectron spectroscopy (XPS) analysis. The electrochemical performance of the integrated pseudo-reference electrode was studied using potassium ferricyanide as a probe molecule and compared against the results obtained from a commercial reference electrode. Its application towards heavy metal ion detection in aqueous solutions was demonstrated.
Presenter bio: Hyoung Jin Cho is a Professor in the Department of Mechanical and Aerospace Engineering at the University of Central Florida. He earned his PhD in Electrical Engineering from the University of Cincinnati in 2002, MS and BS in Inorganic Materials Engineering from Seoul National University in 1991 and 1989, respectively. He worked as Research Engineer at Korea Electronics Technology Institute (KETI) from 1993 to 1997. He was a recipient of NSF CAREER award in 2004. His main research interest is in the development of microscale actuators, sensors and microfluidic components based on MEMS technology.
Hyoung Cho
Short-Term Memory Based Online Learning Framework for Intelligent Sector Selection in IEEE 802.11ad
The IEEE 802.11ad standard operates in the mmWave frequency band and supports data transfer speeds of up to 7 Gbps. However, the signal propagation characteristics in the mmWave band are adverse, and the achievable coverage area is less (around 6m). Hence, the coverage area around the access point (AP) is partitioned into virtual sectors and, AP communicates with stations in each sector by employing directional beamforming for improved coverage. Therefore, at any given instance, the AP serves only one sector, and in general, AP chooses the sectors in a round-robin policy. However, this round-robin based sector selection results in inefficient channel utilization when the traffic generated across the sectors is non-uniform. In this paper, a short-term memory based online learning framework is developed for efficient sector selection by the AP for improving throughput while balancing medium access delay and average queue size at the STAs. The performance of the proposed optimal sector selection policy is compared with the traditional round-robin based sector selection and random sector selection policies. From the performance analysis, it is observed that the proposed sector allocation framework improves the throughput by 51% and 112% when compared with round-robin and random sector allocation policies, respectively.
Presenter bio: Academics: M.Tech and PhD from IIT Madras in Communication Networks Areas of Research: Wireless sensor networks, Cyber Physical Systems, Optical Networks, Embedded Systems
Analysis of smartphone sensor bias from an activity recognition experiment in the wild
A multitude of sensor models are being built-in a variety of smartphone brands. Notably, the accelerometer sensor has been widely used to recognize a variety of human-induced activities leading to growth of human activity recognition(HAR) research. However, diversities in sensor models create heterogeneity in the data which could result in inaccurate recognition accuracies. Most works of experiment in the wild handling diverse set of mobile devices do not pay attention to this issue. In this paper, we report on the bias observed across 13 different accelerometer models spanning 26 different phones. A continuous sensing application was deployed on all phones over a period of 45 days to empirically assess students' behavioral patterns (physical activity patterns). Sensor bias analysis revealed One Plus 6, Poco F1 to have the lowest bias and Samsung S4 the largest bias. Except for Samsung S4(11%), Honor 7(6%)and Nexus 5(6%), all other phones showed less than 3% average deviation from the actual value which is 9.8m/s2. Using the bias ranges observed, we classify the entire unlabelled data with accuracy close to 98%. We believe our results will provide a baseline information before using the phones for continuous sensing applications in general and human activity monitoring in particular.
An IoT-based Discrete Time Markov Chain Model for Analysis and Prediction of Indoor Air Quality Index
Humans generally spend most of their time indoors, therefore, having good Indoor Air Quality (IAQ) and its real time information is critical for maintaining human health and productivity. According to United States Environmental Protection Agency, indoor air even in centrally air-conditioned buildings is several times more polluted than outdoor air, primarily due to change in occupancy pattern, old or ill maintained ventilation systems, and cracks in buildings. In this work, we have proposed an Internet of Things (IoT) based Discrete Time Markov Chain (DTMC) model for analysis and forecasting of IAQ. The IoT architecture used for collecting IAQ data consists of sensing nodes deployed in different rooms of the University building. This sensed data is transferred and stored in IoT cloud and used to generate the IAQ state transition matrix and compute return periods for each state. The predicted and actual return periods have been compared and the accuracy of the proposed model is found to be satisfactory with a low average absolute prediction error of 4.75%.
Presenter bio: Dr. Divya Lohani is currently an Assistant Professor in the Department of Computer Science & Engineering at Shiv Nadar University, India. She is a researcher in the field of Internet of Things, Wireless Sensor Networks and Sensor Data Analytics.
Design of Cyanobyte: An Intermediate Representation to Standardize Digital Peripheral Datasheets for Automatic Code Generation
This is a design for a static configuration representing the registers and functions of a hardware peripheral. This allows for datasheet content to be processed directly in a machine-readable format. This can be used to generate drivers for the peripheral on any variety of hardware and software platforms without manual work in building every driver.
Presenter bio: Nick Felker is a developer programs engineer working on IoT and Assistant at Google. He graduated with a bachelors in Electrical & Computer Engineering from Rowan University.
Provisioning IEEE Smart Transducer Standards (P21451.1) to Include Health Metrics via HEDS
The IEEE 1451.X family of Smart Transducer Standards have been developed over a period of many years. One of the most innovative, integrating elements of the standards is the transducer electronic data sheet (TEDS). This structure was clearly defined in 1451.4-2004. Since that time, the concept has been proposed as a means to describe many other system attributes and operating parameters including components, algorithms, subsystems, and others. System health is another area that has been explored. In our latest work on revisions to P21451.1, we have added the health electronic data sheet (HEDS). While the final location of the HEDS may be incorporated within a sister standard, it is important to have HEDS available to system developers and integrators. HEDS can add additional confidence in the performance of system transducers, which include both sensors and actuators.
Presenter bio: John L. Schmalzel (Ph.D. '80, Kansas State University) is on a two-year IPA assignment from Rowan University to NASA's John C. Stennis Space Center where he works in the Integrated Systems Health Management (ISHM) group in support of rocket engine testing. His current interests primarily focus on the core ISHM technologies of anomaly detection and application of smart sensors to create intelligent sensors containing embedded ISHM capabilities. He has been with Rowan University since joining in 1995 to serve as the founding chair for the ECE program. His previous positions include service as a faculty member at The University of Texas at San Antonio ('84-'95) in the EE Department and as a Captain in the US Army ('80-'84). Within IMS, he has served as VP-Membership and VP-Conferences, and currently is a member of the AdCom. He chaired Sensors Applications Symposium 2006 and 2007.
Development of Non-contact Composite Temperature Sensing (CTS) for photothermal Real-time quantitative PCR Device
Real-time quantitative PCR device was emerging to be an essential platform in precision medicine. Novel photothermal conversion based heating mechanism was reported recently that revolutionize the system to an unprecedented speed and energy efficiency. Under such speed, precision temperature sensing became challenge as the true temperature in each reagent was difficult to be accurately estimated. We developed a new photothermal heating system with thermocouple, metal tape, and thermal conducting rubber serving as a composite temperature sensor (CTS). The temperature of the reagent was also directly detected by thermocouple. The results showed that the CTS with copper tape was more sensitive and stable than that of other groups. A suitable dimension was determined to be 4 x 3 mm for the copper tape with 1 mm thickness of the thermal conducting rubber. The correlation between the two measurements were determined to be 96.2 %. In the future, the CTS outside the tube could be applied to estimate the actual temperature of the reagent in the tube in further PCR applications.
A Novel Experimental Study to Enhance the Attentional State using EEG Signals
In this paper, we propose a simple low-complex classification framework for the cognitive enhancement with the sustained attention stimuli using Electroencephalography (EEG) signals. The visual stimuli comprise of four face images: two happy (one male and one female) and two unhappy (one male and one female). The neuronal response is decoded using a combination of discrete wavelet transform (DWT) and ensemble classifier. The features are extracted by decomposition of recorded EEG signals using Daubechies wavelet filter (db4) and used the statistical methods such as the absolute mean value, power, and standard deviation for classification. The proposed methodology is validated on in-house recorded visual attention EEG (VA-EEG) dataset using six subjects (three males, three females) and evaluated the performance on six binary combinations of facial stimuli. The performance results show that the binary combination of male happy (MH) and female happy (FH) facial stimuli aids in cognitive enhancement for the people suffering from cognitive symptoms. The proposed low-complex feature extraction classification framework obtained a mean classification accuracy (CA) and a mean kappa value of 86.58\% and 0.72, respectively.
Presenter bio: Academics: M.Tech and PhD from IIT Madras in Communication Networks Areas of Research: Wireless sensor networks, Cyber Physical Systems, Optical Networks, Embedded Systems
Exploitation of precise timing capabilities of single board computer for transcranial magnetic stimulation
An early Alzheimer Disease (AD) diagnosis is fundamental for maximizing the effectiveness of treatment administration. Unfortunately, distinguish AD from other neurodegenerative dementias, such as Frontotemporal Dementia (FTD) is not trivial. Transcranial Magnetic Stimulation (TMS) emerged as an effective non-invasive, easy to apply and not time-consuming solution. TMS-based techniques generally require expensive ad hoc clinical equipment that suffer from poor flexibility and user friendliness in defining the specific diagnostic protocol. In this work, a low-cost BeagleBone Black single board computer (BBB-SBC) has been used to implement all the functionalities required to manage a TMS-based instrument. Timeliness of signal generation is guaranteed by dedicated programmable real-time units hosted by the BBB-SBC System on Chip. Web-based interface, complemented by IoT-like features, provide a high degree of versatility and permit the execution of many different diagnostic protocols. In particular, the experimental validation confirms timing error in the sub-microsecond range, more than enough for the considered application.
Presenter bio: Alessandro Depari received the M.Sc. Degree in "Electronics Engineering" and the Ph.D. Degree in "Electronic Instrumentation" from the University of Brescia, Brescia, Italy, in 2002 and 2006, respectively. Since 2007, he has been Assistant Professor and, since 2017, Associate Professor with the Department of Information Engineering, University of Brescia, in the field of electrical and electronic measurements. His research activity includes: sensor signal conditioning and processing, in particular concerning chemical sensors for artificial olfactory systems; embedded systems based on microcontrollers, DSPs and FPGAs; the development of smart sensors and sensor networks for distributed measurement with industrial communication systems; design of methods and digital electronic circuits for numeric measurement instrumentation; design and development of systems for m Health (mobile health) and IoT (Internet of Things) applications.
A Radial Basis Function Technique for the Early Detection and Measurement of Hip Implant Loosening
Hip implants are extremely common procedures that are performed to relieve pain and restore function to the hip joint. Historically, most implants would last for the duration of the patient's life, approximately 15 years after implantation; however, due to medical advancements patients are living longer and outliving the life of their implant. The most common cause for implant failure is the failure of the bond between the implant and the bone, occurring due to the translation or rotation of the implant. Once an implant has failed, a very invasive, costly, and painful revision surgery is required. There is significant advantage in the early detection of implant loosening. This paper presents a radial basis function based image processing technique to detect minute 3-D rotations of a hip implant from 2-D X-ray images. Comparing these rotations for a particular hip implant over time can alert orthopedic surgeons of trends that might lead to impending gross loosening of the implant and enable early correction.
Presenter bio: Shreekanth Mandayam is Professor of Electrical & Computer Engineering (ECE) at Rowan University. Prior to joining the faculty at Rowan in 1997, Dr. Mandayam was a Research Associate and Assistant Professor in the Department of Electrical and Computer Engineering at Iowa State University , Ames, Iowa. Dr. Mandayam received his Bachelor of Engineering degree in Electronics Engineering in 1990 from Bangalore University, India. He received his M.S. degree (Electrical Engineering) in 1993 and his Ph.D. degree in Electrical Engineering (Communications and Signal Processing) in 1996 from Iowa State University. Shreekanth Mandayam's research expertise is in the area of imaging, image processing, multi-sensor data fusion, advanced visualization and virtual reality. Dr. Mandayam is a Senior Member of the IEEE and a member of the American Society for Engineering Education (ASEE).
Shreekanth Mandayam
S11 Calibration Method for a Coaxial-loaded Cut-off Circular Waveguide using SOM Termination
In this study, termination with conditions referred to as short, open and reference material (SOM) was used for S11 calibration at the front of samples with a coaxial-feed-type cut-off circular waveguide as a preliminary step for dielectric measurement in liquids. With pure water as the reference material, the S11 value for the jig at the front of the sample was then calibrated in the frequency band of 0.50 to 3.0 GHz using a VNA (vector network analyzer) via the proposed method. Next, S11 at the front surface of the sample was measured with various liquids in the jig after calibration. The results showed close correspondence of estimated values for each method, thereby indicating the validity of the proposed S11 calibration approach.
Presenter bio: Kouji Shibata was born in Shizuoka, Japan, in 1970. After graduating from the Engineering Department at the Kanazawa Institute of Technology in Ishikawa, Japan, in 1993, he took up employment with SPC Electronics Corp. in Chofu, Japan. His roles there included designing and developing passive circuits such as filters, couplers, diplexers and antennas in the microwave/ millimeter band. In 2001 and 2004, respectively, he completed the Master's and Doctoral Programs of Aoyama Gakuin University Graduate School in Tokyo, Japan, and holds Doctor of Engineering status. In 2004 he took up a position as a lecturer at Hachinohe Institute of Technology in Japan and became an associate professor in 2015. He is currently engaged in research on electrical constant measurement for passive components in the high-frequency band
Kouji Shibata
Mechanical Requirements for a Smart Inhaler Product
Asthma and other respiratory diseases are a significant problem for the New Zealand health sector. There are several factors that make asthma treatment difficult and these have led to the development of smart inhaler devices. However, the devices currently available on the market do not address the core problems with asthma treatment and so a new smart inhaler product is being developed. The smart inhaler product uses an acoustic sensor to detect inhalation flow rate and a haptic motor to provide vibration feedback. To avoid the vibration feedback interfering with the acoustic sensing, vibration damping methods have been employed. Passive damping using viscoelastic material was chosen as it is low-cost and compact. However, results show that these methods are not effective enough to reduce the magnitude of the vibration noise below the magnitude of the inhalation signal using the current acoustic sensor (a microphone). It is possible that effective damping could be achieved using a transducer as the acoustic sensor. However, the transducer must be integrated with the smart inhaler product and likely isolated in some way. Following this, further testing would be required.
Presenter bio: Baden is a PhD student studying at Massey University, New Zealand. His doctoral research focuses on automating grape yield estimation in vineyards using 3D camera technology. In addition to these studies, Baden is also involved in research on indoor localization using visible light, acoustic imaging using active and passive arrays, dense urban IoT sensor networks for air quality monitoring, and robot design and locomotion strategies.
Phase based Time Resolved Reflectance Spectroscopy using Time-of-Flight Camera for Fruit Quality Monitoring
Time-resolved phase-based reflectance spectroscopy has been used for the nondestructive assessment of the internal quality of the fruits in this work. The time taken by the incident light to backscatter and reach the detector placed at a distance from the camera relates to the penetration depth or mean free path of the light into the apples. The mean free path changes due to chemical pigmentation changes in apples during storage and ripening. Therefore, the scattering characteristics changes. The phase changes in the backscattered light are correlated to the stage of ripening or storage. The degree of linear polarization is used to quantify the changes in the reflected phase with increasing storage time. The proposed optical instrumentation is simple in construction and computationally less intensive and therefore, provides an affordable solution for a handheld compact fruit monitoring device.
Presenter bio: I am an Assistant Professor with Indian Institute of Technology, Delhi (IIT Delhi) in the Electrical Engineering Department. Prior to this I have spent a year as postdoctoral researcher with Electronic Instrumentation Laboratory, Technical University of Delft, The Netherlands. I received my Ph.D. in 2011 from Technical University of Delft, The Netherlands. I was a full time resident of IMEC from 2007 to 2011 during my Ph.D. Between 2003 and 2005, I worked in the Philips Institute of medical information, Aachen, Germany as a research assistant in detection and analysis of bio-signals. My research interests lie in the areas of solid state imaging, CMOS Image sensors, Bio-inspired vision systems, Analog/Digital circuit design, Optoelectronics and Machine vision.
Model-based Calibration of a Magnetic Induction Spectroscopy System for Absolute Conductivity Measurement
Magnetic induction spectroscopy (MIS) is a measurement technique by which the impedance spectra of an object can be determined by inducing eddy-currents in the object, then detecting the resultant magnetic field. It is a powerful approach as it entirely non-contact, but suffers the major drawback of only being able to return relative impedance contingent on the shape of the test object, rather than independent and physically meaningful absolute measurements. In this paper, we address this shortfall by demonstrating a simulation-derived procedure to obtain true conductivities from MIS measurements. We determine calibration constants by using saline solutions of different sizes and conductivities and find an approximate accuracy of +/- 65 mS/m. We further employ this procedure to find approximate absolute conductivities of a baking potato, braeburn apple and conference pear over part of their beta-dispersion region between 100 kHz and 10 MHz.
Presenter bio: Michael O'Toole received an MEng (Hons) in Integrated Engineering from the University of Reading in 2006 and a PhD from Loughborough University in 2011. He has been at the University of Manchester since 2011, originally a research associate, then Leverhulme Trust Early Career research fellow and now Lecturer. His primary research interest is in the design of new electromagnetic and induction based sensors for impedance spectroscopy. His work focuses on magnetic induction systems for inspection of food, healthcare and medical applications, and resource recovery of non-ferrous metals.
Development of a Hand-held 3D Scanning Acoustic Camera
Traditionally, acoustic cameras have used a fixed 2D near-field scanning plane for localising noise sources. This has been shown to cause errors in captured data and restricts acoustic observation to a fixed viewing position. With the recent popularisation of consumer 3D depth cameras and open source 3D reconstruction algorithms, researchers have begun investigating their utility in generating 3D near-field evaluation surfaces to increase beamforming accuracy. This paper presents the design and development of a portable hand held 3D scanning microphone array based on comprehensive simulations. Steps taken to improve the camera towards real-time construction of 3D acoustic models are also discussed.
Presenter bio: Baden is a PhD student studying at Massey University, New Zealand. His doctoral research focuses on automating grape yield estimation in vineyards using 3D camera technology. In addition to these studies, Baden is also involved in research on indoor localization using visible light, acoustic imaging using active and passive arrays, dense urban IoT sensor networks for air quality monitoring, and robot design and locomotion strategies.
Indium Tin Oxide Films Based pH Concentration Sensor Fabrication and Verification Using Pulsed UV Laser Patterning Technology
A pH sensor with sensitive material, laser-patterned indium tin oxide films on a glass substrate, was developed in this study. A 10 μL droplet prepared by buffer solution was injected to the sensing area (i.e. designed interdigitated ITO electrode), and the resistance variation of the electrode was recorded to build up the model of pH concentration detection. The result indicated that the interdigitated electrode pattern with a line width of 500 µm and a line pitch of 100 µm was the most sensitive under various pH conditions. In addition, for the acid solution and alkali solution, when the interdigitated electrode pattern with a line width of 100 µm and a line pitch of 500 µm was more sensitive in the acid environment, whereas the interdigitated electrode pattern with a line width of 500 µm and a line pitch of 100 µm can be applied in the alkaline environment due to its great sensitivity. During the pH value monitoring, the response time of resistance changing was below 3 seconds to reach the steady-state resistance in all various pH conditions, and the correlation coefficient was over 0.9 validated by a quadratic regression in both acid and alkaline sensing verification.
Presenter bio: Dr. Wen-Tse Hsiao, received his Ph.D degree in Mechatronics engineering from National Changhua University of Education in 2009. He joined the Instrument Technology Research Center (ITRC) of National Applied Research Laboratories, Taiwan in 2009, where he is now an Research Fellow. His research interests include laser machining system design, laser micro/nano machining, surface treatment, laser-matter interactions, laser interferometer measurement, precision motion control, and human machine interface programming design. He is now the reviewers of Appl. Surf. Sci. (SCI journal), Int. J. Adv. Manuf. Tech. (SCI journal), JNN (SCI journal), J Appl. Polym. Sci. (SCI journal). Applied Physics B: Lasers and Optics. (SCI journal), Optics and Lasers in Engineering. (SCI journal) Solar Energy Materials & Solar Cells. (SCI journal) Precision Engineering. (SCI journal) Optics & Laser Technology. (SCI journal) Journal of Materials Processing Technology. (SCI journal)
Evaluation of a Digital Converter for Linear and Nonlinear Temperature Sensors
A simple digital converter suitable for linear and nonlinear temperature sensor is proposed in this paper. This digital converter uses easily-available analog components and a digital time measurement unit in its architecture. The converter uses the same hardware architecture for both linear and nonlinear sensors. Novel linearization approaches are used to extract a linear output in case of nonlinear sensors. The circuit also provides other desired features like independence from power-supply drifts and many Op-amp non-idealities. The methodology of the converter and design/linearization approaches used are explained in the paper. The possible error sources of the circuit are also analyzed theoretically in this paper. These studies show that the developed converter can serve as an accurate and efficient interface for temperature sensors. The performance of the digital converter is verified by conducting various (simulation and experimental) studies on linear and nonlinear temperature sensors. The linearization capability of the scheme, in the case of nonlinear sensors is also clearly illustrated in the paper.
Presenter bio: Research scholar
An Efficient Multi-AMR Control Framework for Parcel Sorting Centers
With the growth of eCommerce activities, there is an urgent need for robust logistics solutions to ensure mass and speedy parcel delivery. Upon the parcels are collected, one of the main issues is how to efficiently sort a huge number of parcels at a parcel sorting center to reduce the delivery cost. To achieve this, autonomous mobile robots (AMRs) can be adopted for automating the parcel sorting. This paper studies the multi-AMR control framework for automated parcel sorting centers. Traditionally, multi-mobile robot control frameworks in indoor environment typically employ a centralized approach. However, these methods may not scale, as they make little use of the computation ability of the robots. On the other hand, global information might not be well utilized in most of the distributed approaches. Therefore, this paper proposes a hybrid framework, and its decision modules are distributed between the server and the clients. On the other hand, this paper examines how road regulation affects the throughput of the parcel sorting center. The experimental results show that the proposed framework can achieve good performance and it can even be further improved by designing a reasonable road regulation.
Presenter bio: Chee Henn is a postgraduate student at University Tunku Abdul Rahman (UTAR). He received a Bachelor of Computer Science (Hons) Degree with Distinction from the same university in 2019. His current research area is multi-robot systems for fast and high-density robots.
LifeCount: A Device-free CSI-based Human Counting Solution for Emergency Building Evacuations
During large scale building evacuations, it is difficult to ascertain how many people have been evacuated safely. To assist in the rescue effort, indoor counting solutions can provide emergency personal with the number of people who have evacuated the building, and from which floors. LifeCount implements a novel two stage neural network algorithm to accurately count the number of people passing through a hallway. Experimental results show that LifeCount can attain a zero counting error accuracy of 96.9%.
Presenter bio: Baden is a PhD student studying at Massey University, New Zealand. His doctoral research focuses on automating grape yield estimation in vineyards using 3D camera technology. In addition to these studies, Baden is also involved in research on indoor localization using visible light, acoustic imaging using active and passive arrays, dense urban IoT sensor networks for air quality monitoring, and robot design and locomotion strategies.
Fingerprint-Based Visible Light Positioning using Multiple Photodiode Receiver
Visible Light Positioning (VLP) is a promising indoor localization method as it provides high positioning accuracy and allows for leveraging the existing lighting infrastructure. Photodiode (PD)- based receiver is a commonly used tag for VLP. However a tag employing single PD requires three or more luminaires to be visible. This paper presents a VLP system that uses a custom made tag utilizing multiple PDs. It applies Received Signal Strength (RSS)-based fingerprinting using Weighted k-Nearest Neighbor (WkNN) algorithm for localization. Experimental results show that it is possible to localize using less than three luminaires with high accuracy. Manhattan and Matusita distance metrics are found to provide lower localization accuracy than the Euclidean metric for the WkNN algorithm.
Presenter bio: Baden is a PhD student studying at Massey University, New Zealand. His doctoral research focuses on automating grape yield estimation in vineyards using 3D camera technology. In addition to these studies, Baden is also involved in research on indoor localization using visible light, acoustic imaging using active and passive arrays, dense urban IoT sensor networks for air quality monitoring, and robot design and locomotion strategies.
On the use of LoRaWAN for the Internet of Intelligent Vehicles in Smart City scenarios
Automotive world is changing with the introduction of Intelligent Vehicles. Today, an increasing number of vehicles may send data to the Internet, helping car manufacturer to implement the Industry 4.0 paradigm. The collected data during the lifetime of the vehicles can be used to improve both product and production facilities. Moreover, the availability of additional data coming from onboard sensors could be used to obtain information about the environment surrounding the vehicle. The Smart City scenario includes an extraordinary number of new sensors (in urban area) and vehicles can be thought as additional mobile sensors. This paper describes the prototype of a Vehicle-to-Cloud interface with OBD-II (On Board Diagnostic) communication, 3G/4G connectivity, and LoRaWAN [used as backup channel]. LoRaWAN infrastructures are largely diffused in Smart Cities and they can provide a suitable alternative to cover some areas when 3G/4G fails. Last, considering the Smart City scenarios, this paper discusses the application constrains and design directions to achieve a correct integration between LoRaWAN infrastructure and the Internet of Intelligent Vehicles.
Presenter bio: Alessandro Depari received the M.Sc. Degree in "Electronics Engineering" and the Ph.D. Degree in "Electronic Instrumentation" from the University of Brescia, Brescia, Italy, in 2002 and 2006, respectively. Since 2007, he has been Assistant Professor and, since 2017, Associate Professor with the Department of Information Engineering, University of Brescia, in the field of electrical and electronic measurements. His research activity includes: sensor signal conditioning and processing, in particular concerning chemical sensors for artificial olfactory systems; embedded systems based on microcontrollers, DSPs and FPGAs; the development of smart sensors and sensor networks for distributed measurement with industrial communication systems; design of methods and digital electronic circuits for numeric measurement instrumentation; design and development of systems for m Health (mobile health) and IoT (Internet of Things) applications.
A measurement system to investigate dielectric properties of flexible substrates for sensing applications
The rapid prototyping of flexible and low-cost sensors is becoming a real need, especially for research purposes and applications requiring customizable and shapeable sensing solutions. In this paper a dedicated measurement system aimed at investigating the dielectric properties of different flexible materials is presented, aimed at the development of a low-cost, flexible, customizable, inkjet printed pressure sensor. This research activity is developed within an INTERREG Italy - Malta Project, aimed at developing an innovative framework of assistive technology supporting elderly and frail people.

TRR: Recent Results

Room: Foyer 2
An insole 3D force sensor for gait analysis
A 3D force sensor, implemented in an inlay sole, is presented. The wireless sensor system provides real-time force data while walking, which may help optimize orthopedic innersole design.
Long-Term Monitoring System for Marine Lobster Aquaculture in Vietnam
In this paper, a prototype monitoring application based on the Internet of Things network is proposed for lobster farming in Xuan Dai Bay, Vietnam. The platform is leveraged by Android Things Operating System installed on a Raspberry PI 3B board. In order to prolong the use of sensors in marine environment, a pump system is also proposed. Our works aim to provide real-time sensory data that enables farmers to quickly react to emerging issues and critical changes in lobster farming.

TT@S: TIM @ SAS 2020

Room: Foyer 3
Blind Separation of Doppler Human Gesture Signals Based on Continuous-Wave Radar Sensors
Recently, progresses have been made in hand gesture recognition based on Doppler radar sensors. However, it remains a technical challenge to avoid the interfering human motions. An example is detecting hand gestures in the presence of random body movements. In this paper, we propose a solution based on a ingle-input Multipleoutput frontend and a blind motion separation algorithm. Assisted by an additional receiving channel, Doppler signals caused by different motions can be separated by extending the algorithm originally developed for separating human voices. The experimental separation of hand gestures from interfering movements verified the effectiveness of this approach. The proposed solution can be potentially used in novel applications such as human gait and gesture recognitions.

Tuesday, March 10 14:00 - 15:40

TA2: MEMS and Nano-Sensors

Room A
Chairs: Behraad Bahreyni, Serge Demidenko
14:00 Fabrication of Tin Oxide Based Gas Sensor in Ethanol Gas Sensing
In this study, the tin oxide (SnO2) based gas sensor with micro heater embedded (working temperature lower than 100 °C) was developed, and the ethanol-sensing characteristic was also discussed. The mico heater presented a fast, stable heating source while gas sensing process. The result indicated that the gas sensor had more than 88% in sensitivity regarding to the 10 ppm and 30 ppm ethanol gas sensing.
Presenter bio: Dr. Wen-Tse Hsiao, received his Ph.D degree in Mechatronics engineering from National Changhua University of Education in 2009. He joined the Instrument Technology Research Center (ITRC) of National Applied Research Laboratories, Taiwan in 2009, where he is now an Research Fellow. His research interests include laser machining system design, laser micro/nano machining, surface treatment, laser-matter interactions, laser interferometer measurement, precision motion control, and human machine interface programming design. He is now the reviewers of Appl. Surf. Sci. (SCI journal), Int. J. Adv. Manuf. Tech. (SCI journal), JNN (SCI journal), J Appl. Polym. Sci. (SCI journal). Applied Physics B: Lasers and Optics. (SCI journal), Optics and Lasers in Engineering. (SCI journal) Solar Energy Materials & Solar Cells. (SCI journal) Precision Engineering. (SCI journal) Optics & Laser Technology. (SCI journal) Journal of Materials Processing Technology. (SCI journal)
14:20 Dual-axis Lorentz Force MEMS Magnetometer
This paper presents a dual axis MEMS magnetometer, utilizing two DOF torsional gyroscope structure. The designed structure is a torsional resonator with two gimbals. Single structure is capable of detecting magnetic field in two directions. First and Second mode of vibration are the desired mode of operation for detecting the magnetic field in x and y directions. The first and second mode of vibrations is at 107 kHz and 187 kHz respectively. The device is tested for its Lorentz force transduction using MSA-500 in the presence of magnetic field generated using permanent magnet for both the axes at atmospheric pressure. The fabrication process is based on anodic bonding (<400°C) of a borofloat glass wafer and double side polished (DSP) Si wafer, which enables the passivation between Gold loop and Silicon.
14:40 Fail-Operational Shock Detection and Correction of MEMS-based Micro-Scanning LiDAR Systems
Highly automated or autonomous vehicles will be dependent on systems that have to perceive the environment to make valid decisions during the driving cycle. One of the key enablers for autonomous and highly automated vehicles will be Light Detection And Ranging (LiDAR) technology. A Micro-Electro-Mechanical System (MEMS) based Micro-Scanning LiDAR is able to detect obstacles in a predefined Field-of-View (FoV). The point cloud stability of the scanned FoV is mandatory to be able to make a valid point where the obstacle is located in the scenery. Due to the fact that massive shocks can occur arbitrarily to the system, it is necessary to be able to detect and correct them as fast as possible that point cloud stability can be recovered as fast as possible. In this paper, we introduce a novel system architecture that enables a fast shock detection and correction of phase and frequency for MEMS-based Micro-Scanning LiDAR Systems. Our novel introduced fail-operational detection and correction system architecture was implemented in a 1D MEMS-based Micro-Scanning LiDAR FPGA platform to prove its feasibility and for performance evaluation.
Presenter bio: Philipp Stelzer is a PhD student of the Institute of Technical Informatics at Graz University of Technology. His research areas are Hardware/Software Codesign and Embedded Automotive Systems. He is especially focusing on smart mobility, safety and reliability of MEMS-based LIDAR Systems.
15:00 Design, Modeling and Simulation of MEMS Resonator for Humidity Sensor Application
Humidity sensor plays an important role in our daily life as well as industries. Microelectromechanical systems (MEMS) humidity sensors are well developed and have high efficiency. However, it has problems such as high damping coefficient and low sensitivity. Thus, this research will focus on effect of damping coefficient on MEMS resonator. Mathematical modeling and optimize it through finite element analysis simulation to investigate device characteristics such as spring constant, resonant frequency and air damping. The device will be designed based on standard PolyMUMPs process technology with electrothermal actuation method and capacitive sensing technique to measure the output voltage. The effect of changes in length and width of the beam on spring constant and resonance frequency are investigated. The spring constant is found to be decreasing when the length of the beam increases and increasing when the width of the beam increases. On the other hand, resonance frequency is found to be decreasing when the length increases and increasing when the width of the beam increases. It is also found that the damping coefficient decreases where the number of holes and radii of holes increases. Analytical and simulation results of frequencies showed good agreement within percentage difference of 0.04 -1.23%
Presenter bio: I am Ashaashvini Mutharpavalar, currently pursing Master of Science in Electrical and Electronics Engineering at Universiti Teknologi PETRONAS and area of focus is microelectromechanical system (MEMS) sensor design and microfabrication technology based on PolyMUMPS technologies. I have completed my Bachelors of Degree in Electrical and Electronics Engineering from Universiti Teknologi PETRONAS with Major in Electronics and Devices.
15:20 Effect of Oscillator Phase Noise on Synchronous Demodulation Measurement Systems for Sensing Applications
Synchronous demodulation is a well-established technique for precise definition and reduction of noise bandwidth in various applications. Typically the noise added by the nonlinear elements such as reference oscillator and demodulator is assumed to be negligible. However, with the scaling of the dimensions of micro-sensors, the signals from these sensors generally tend to become weaker. Therefore, there is a need to study the so far neglected noise contributions from the components of the synchronous demodulator. In this paper, we focus on the significance of the phase noise of the reference oscillator on the system performance. A detailed analytical model is developed to investigate the nonlinear interaction. It is shown that the phase noise of the reference oscillator can significantly contribute to the output noise depending on the phase difference between the reference and measured signals. Close-to-resonance phase noise components produce low-frequency components at the output. However, our experimentally tested analysis indicates that the effect of phase noise on the output signal can be eliminated by complete compensation of the phase shift between reference and measured signals. This study applies to the design of low cost, high-performance measurement systems for high precision sensor applications.

TB2: Remote Sensing and Sensor Networks

Room B
Chair: Liam Marsh
14:00 Multi-Purpose Marine Sensor Buoy
Studies in the marine environment require devices to gather data in remote locations. These devices commonly use some form of data logging that require the user to retrieve the device after a period of time. This is not only a large effort, but often presents an unnecessary delay and risk in the event that the device malfunctions, gets lost or if the data collected is redundant. We propose a modular, remote and autonomous sensing solution that enables multi-sensor readout and wireless data communication, providing scientists with real time data. This sensor buoy is comprised of a primary module which is a data logger that connects to an online server via mobile network. The secondary portion is the hub that connects to an array of sensors that are customized to the needs of the marine scientists.
Presenter bio: Alexander Przybysz is a research engineer for the Sensing, Magnetism and Microsystems (SMM) group in King Abdullah University of Science and Technology, Saudi Arabia. He graduated from Stellenbosch University in South Africa and holds a Bachelors degree in electrical and electronic engineering. He has worked in the mining industry, where he gained industrial experience. His research interests are in the area of sensors and sensor applications with a focus on research translation and innovation.
14:20 Long Range Nuclear Radiation Monitoring System using LPWAN Technology
A nuclear radiation monitoring system with a transmission range of 10 km is developed and validated. The designed system can be used for remote monitoring of nuclear radiation and collecting the data at ground stations located far apart from active site. It can be used to interface any wireless nuclear radiation sensor and has ~10 years of operating lifetime. The transceiver is implemented using novel Low Power Wide Area Network (LPWAN) technology which is specially designed for applications requiring long range transmissions at low data rate. The experimental testing of designed system shows that dose rate information can be transmitted upto 10 km in rural and 7 km in urban environment. Moreover, there is possibility of creating a network where number of different sensors located geographically apart can transmit to the same base station, thus providing simultaneous data analysis.
14:40 CIG based Stress Identification Method for Maize Crop using UAV based Remote Sensing
The health and yield of crops depend on the use of water, nutrients, and fertilizers. Due to climatic changes and reduction in rainfall, farmers are relying on groundwater for irrigation, which should be used optimally. The use of water and other agronomic inputs can be optimized by monitoring the health of crops and soil. Usually, it is done by manual observation, which is labor-intensive and time-consuming. In this paper, we propose Chlorophyll Index Green (CIG) vegetative index-based method for monitoring the crop health using near-infrared, green, and red band images acquired using a multispectral camera mounted on Unmanned Ariel Vehicle (UAV). The proposed method clearly classifies the water-stressed area of the field and helps in optimizing the irrigation process and monitoring the crop-health.
Presenter bio: Academics: M.Tech and PhD from IIT Madras in Communication Networks Areas of Research: Wireless sensor networks, Cyber Physical Systems, Optical Networks, Embedded Systems
15:00 Solving Surveillance Coverage Demand Based on Dynamic Programming
The recent advances on visual sensor placement attempt to increase the coverage area and/or decrease the installation cost of a visual sensor array. Typical visual sensor planning approaches install visual sensors without stress on coverage demand. This paper addresses the visual sensor coverage prioritization based on risk mapping and dynamic programming. The optimization technique utilizes the outcomes of a prior routine which highlights the significance regions, then perform the sensor planning accordingly. A dynamic programming approach is used to compute a near optimal coverage solution. The results show the reliability of the visual sensor planning with risk maps using dynamic programming.
Presenter bio: Altahir A. Altahir is with the Department of Electrical and Electronics Engineering, Universiti Teknologi PETRONAS (UTP).
Altahir Abdalla Altahir
15:20 Evaluation of a Bespoke Antarctic Meteorite Detection System in Polar Operating Conditions
A bespoke detection system has been constructed for the purpose of identifying iron-rich meteorites which are suspected to exist beneath the ice surface in regions of Antarctica. This system has been specially hardened to the hostile polar environment, and is capable of searching a desired target space of 1 km^2 in approximately 11 hours, maintaining a target speed of 15 km/h. The performance of the detector has been validated through structured testing in the High Arctic. A test site was constructed in Ny-Alesund, Svalbard, which contained 13 meteorite surrogates buried at depths between 0 cm (flush with the ground level) and 40 cm. Representative results for detector performance in typical Antarctic operating conditions are reported.The system demonstrates the ability to reliably detect targets as deep as 20 cm using real-time processing, and shows the potential for identifying targets as deep as 30-40 cm with more advanced post-processing, or improvements in the real-time discrimination algorithm.
Presenter bio: Dr Liam Marsh is a Lecturer in Embedded Systems at the University of Manchester, UK. Liam gained a first class MEng degree (with honours) in Electrical Engineering and Electronics from the University of Manchester Institute of Science and Technology (UMIST) in 2007. He studied for his PhD ‘Electromagnetic Tomography and People Screening’ at the University of Manchester. In 2011 he took up the post of Research Associate at the University of Manchester. In 2018 Liam was appointed as a Lecturer in Embedded Systems in the Department of Electrical and Electronic Engineering at the University of Manchester. Liam’s research includes many aspects of metal detection and characterisation, magnetic induction spectroscopy/tomography, signal processing, remote sensing and security. In the past he has also worked in bio-impedance, an area in which he maintains an active interest. In recent years Liam's research has focused on the detection and characterisation of landmines (including explosive remnants of warfare), and the development of magnetic sensing systems for polar regions.

Tuesday, March 10 16:10 - 17:50

TA3: Innovative Sensing

Room A
Chair: Hyoung Cho
16:10 Towards Plastic Optical Fiber Magnetic Field Sensors exploiting Magnetic Fluids and Multimode SPR-POF platforms
We have studied how the optical property of a multimode plastic optical fiber (POF) can be used to monitor a magnetic field, exploiting a Surface Plasmon Resonance (SPR) POF sensor. To realize the sensor system, we have deposited a magnetic fluid on the cladding of a multimode POF connected in input to an SPR D-shaped POF sensor platform. The target magnetic field exerts a bending force on the POF covered with magnetic fluid, thus producing a change of light in input to the SPR POF sensor and modifying the SPR phenomenon. This work presents preliminary experimental results to demonstrate that this novel principle of sensing can be used for highly sensitive measuring of the magnetic field. Preliminary results have demonstrated the viability of this approach showing the possibility to obtain higher sensitivities, compared with the state of the art on SPR optical fiber magnetic field sensors exploiting magnetic fluids.
Presenter bio: Nunzio Cennamo was born in Naples (Italy) in 1975. He received the degree in electronic engineering in 2002 and the Ph.D. degree in electronic engineering in 2005, both from Second University of Naples, Italy. He is professor of electronics at the at the Department of Engineering at University of Campania “Luigi Vanvitelli”, Italy. His research interests include the design and fabrication of optical sensors, chemical sensors, biosensors and optoelectronic devices. He is author of more than 80 international journal and conference papers and 3 patents. He is a reviewer for several technical journals, Member of the Scientific Committee of International Conferences, and Member of the Editorial Board of International Journals.
16:30 A novel approach to optically distinguish plastics based on fluorescence lifetime measurements
In medical and biological research, fluorescence lifetime measurements and fluorescence lifetime imaging is already a part of the standardized analysis procedures. As first investigations have shown, polymers can be identified using fluorescence lifetime imaging and an evaluation algorithm. Thus, this contribution pursues a novel approach for the direct differentiation of four polymers with fluorescence lifetime imaging. Therefore, the evaluation algorithm is extended to compare several fluorescence lifetime images to prove that a distinction is possible.
Presenter bio: Maximilian Wohlschläger 26 years old Masters degree at TH Rosenheim with the Topic of Optical Identification of Polymers using fluroescence.
16:50 Disposable Sensor Devices Fabricated by Paper Crafting Tools
Disposable sensor device concepts are presented in this work by the use of traditional paper crafting tools and techniques. A printer, a paper-cutting machine, and a laminator were used to create electrodes and hydrophobic barriers. With the combination of different paper crafting techniques, electrodes as well as microfluidic channels could be fabricated. For demonstration, a paper-based microfluidic device with impedance readout is introduced in this study. Laser printed polyester particles from a toner were used as an adhesive interface for a metal film transfer process, resulting in a defined electrode pattern. A cut pattern produced by a paper-cutting machine was used to produce distinct hydrophobic barriers, which constituted microfluidic channels and sample reservoirs. The application was demonstrated by developing flexible interdigitated electrodes (IDEs). The developed IDEs could be used to determine chemical concentrations of electrolyte and glucose. The presented fabrication method is suitable for rapid, low-cost production of flexible disposable sensors from digital files using well-known crafting practices.
Presenter bio: Hyoung Jin Cho is a Professor in the Department of Mechanical and Aerospace Engineering at the University of Central Florida. He earned his PhD in Electrical Engineering from the University of Cincinnati in 2002, MS and BS in Inorganic Materials Engineering from Seoul National University in 1991 and 1989, respectively. He worked as Research Engineer at Korea Electronics Technology Institute (KETI) from 1993 to 1997. He was a recipient of NSF CAREER award in 2004. His main research interest is in the development of microscale actuators, sensors and microfluidic components based on MEMS technology.
Hyoung Cho
17:10 Detecting Powerline Noise with Low-Cost Noise Sensors for Power Outage Mitigation
Relatively low-cost noise sensors can be utilized to detect for aberrant powerline noise, which is an early indicator and warning of potential power reliability and stability issues. By taking preemptive action, power outages may be avoided. Accordingly, the practicality as pertains to the utilization of low-cost noise sensors (segueing to scalability and extensibility) is examined. The strategic placement of sensors at key distribution poles has been examined previously. This paper examines how low-cost noise sensors might provide an opportunity for comprehensive coverage and more likely detection of certain powerline noise aberrations at the "edge" against a set of compiled heuristics.
Presenter bio: Dr. Chan has authored papers for various IEEE conference proceedings, such as IEEE Sensors Applications Symposium (SAS), IEEE Computing and Communication Workshop and Conference (CCWC), IEEE Information Technology, Electronics & Mobile Communication Conference (IEMCON), IEEE International Conference on Information & Communications Technology (ICOIACT), IEEE Future of Information and Communication Conference (FICC), IEEE Future Technologies Conference (FTC), IEEE International Conference on Digital Ecosystems and Technologies (DEST), and IEEE International Conference on Collaborative Computing: Networking, Applications, and Worksharing (CollaborateCom).
17:30 Green Nonlinear Energy Harvester from Vibrations based on Bacterial Cellulose
In this paper, we present a green, biodegradable and eco-friendly energy harvester from wideband mechanical vibrations. The proposed device is based on bacterial cellulose, produced by some bacteria impregnated with ionic liquids, and covered with conducting polymers. Tanks to the mechanoelectrical transduction properties of this composite, power is generated, because of mechanical deformation. Furthermore, in order to increase the generated output voltage and the spectral content of the device, a nonlinear behavior has been implemented. In particular, results and discussion about how this dynamic improves the characteristic of the green energy harvester, compared to a classical resonator are presented. A suitable setup has been conceived, realized, and an experimental campaign has been accomplished obtaining very interesting results.
Presenter bio: Prof. S. Graziani received the M.S. degree in electronic engineering and the Ph.D. degree in electrical engineering from the Universit? di Catania, Catania, Italy, in 1990 and 1994, respectively. Since 1990, he has been with the Dipartimento Elettrico Elettronico e Sistemistico, Universit? di Catania, where he is Associate Professor of electric and electronic measurement and instrumentation. His primary research interests lie in the fields of signal processing, multisensor data fusion, neural network, fuzzy logic, and smart sensors. During his activity, he has co-authored several scientific papers and one book.

TB3: Smart Agriculture and Smart Buildings

Room B
Chair: Baden Parr
16:10 Design of Low Cost Programmable LED Lighting System for Smart Buildings
Smart LED based lighting systems have significant advantages over traditional lighting systems due to their capability of producing tunable light spectra on demand. This type of system will interface seamlessly with any intelligent operating system to optimize lighting either to control the human circadian rhythm (i.e. Human-Computer Interaction), or for more effective machine vision applications (i.e. Machine-to-Machine communication). The main challenge in the design of a multichannel smart lighting system is to produce sufficient intensity uniformity over the illuminated region by all channels, while simultaneously maintaining affordability. This paper outlines the design and fabrication of a programmable LED lighting system to achieve these aims. It presents a nine channel lighting design using low cost discrete LEDs.
16:30 Cloud based Low-Power Long-Range IoT Network for Soil Moisture monitoring in Agriculture
Intervention of sensors and wireless networks has transformed cliched agricultural practices. Internet of Things(IoT) has penetrated various verticals with agriculture being one of them. Application of IoT in agriculture is largely focused on field parameter monitoring and automation, which aims to help farmers increase crop yield. Long-range and low-power devices, convenient installation, and cost-efficiency are the primary factors to be considered for deploying an IoT network in realtime. In this paper, we proposed a low-power long-range IoT network for monitoring of soil moisture. We have selected LoRa as the communication interface, which uses 868MHz ISM band for signal transmission. The soil-moisture sensor and the LoRa nodes are designed in-house. Accuracy of the sensor nodes is tested by placing two nodes in the same sector. All the data collected are stored in the server and is available online.
16:50 Multisensor device for emergency recognition in smart building environment
In case of serious environmental disasters, such as floods, landslides and earthquakes, the first activity that an emergency management system must carry out is the detection, recognition and report of anomalous events. In order to be able to automatically perform these actions, it is necessary to have an efficient sensor network within the building. By means of a well-organized and high-performance sensor network, data from the field are acquired and processed to produce usable information at a high level. The Building Management System (BMS) therefore has information available on the occurrence of an anomalous event, whether caused by natural or intentional causes. The purpose of our work is to design and implement a cost-effective sensorized device, managed by an ATmega microcontroller and equipped with a Wi-Fi module for easier integration, able to detect and report potentially dangerous situations within a building. In particular, fires, floods, gas leaks and earthquakes were considered. Thanks to the integration of data fusion algorithms, the platform is able to recognize and report the occurrence of emergency events within a building.
Presenter bio: Dario Masucci is a Ph.D student in Computer Science and Automation Engineering at “Roma Tre” University, where he received the master degree in 2015. He is working as junior researcher at Models for the Critical Infrastructure Protection Laboratory. His research activities include multi-objective optimization algorithms, Multi-Criteria Decision Making, for energy sustainability and emergency situationmanagement. He is also working with the Energy Division in the ENEA Research Center "La Casaccia", for the anomaly detection in Building Management System.
17:10 Acoustic Identification of Grape Clusters Occluded by Foliage
The performance of a vineyard can be influenced by accurate yield estimations prior to harvest. Traditionally, this is a manual process. However, due to the high labour costs and subjective nature of manual assessments, researchers have been working on automated techniques. Utilising 2D computer vision has shown promising results but is inherently limited due to occlusions. The algorithms can only count grapes that are directly visible. Often this shortcoming is accounted for by using coarse occlusion ratio estimates, which themselves need to be manually determined. As a result, researchers have begun looking at alternative methods of grape detection. Synthetic Aperture Radar (SAR) has been demonstrated as a feasible approach to see grape clusters behind leaves. However, this comes at a significant financial cost. This paper introduces an alternative approach that utilises low frequency ultrasound to detect grape clusters in the presence of foliage occlusion. We demonstrate that such low frequency signals have the ability to propagate through foliage and reflect off grapes behind. Additionally, by agitating the leaves we can analyse the variance of consecutive samples and determine which volumes are likely to belong to grape clusters.
Presenter bio: Baden is a PhD student studying at Massey University, New Zealand. His doctoral research focuses on automating grape yield estimation in vineyards using 3D camera technology. In addition to these studies, Baden is also involved in research on indoor localization using visible light, acoustic imaging using active and passive arrays, dense urban IoT sensor networks for air quality monitoring, and robot design and locomotion strategies.
17:30 Device Free Localization with Capacitive Sensing Floor
Passive indoor positioning has many applications including intrusion detection, fall detection of the elderly, and occupancy sensing to name a few. However, current Device Free Localization (DFL) solutions fall short of the desired accuracy requirements and are difficult to implement in a real-world scenario. This research investigates the use of a capacitive floor-based sensing solution, which can simultaneously detect multiple footsteps of a subject. The developed sensing floor prototype achieved a median positioning error of 13.5 mm and a median angular accuracy of 10.4°.
Presenter bio: Baden is a PhD student studying at Massey University, New Zealand. His doctoral research focuses on automating grape yield estimation in vineyards using 3D camera technology. In addition to these studies, Baden is also involved in research on indoor localization using visible light, acoustic imaging using active and passive arrays, dense urban IoT sensor networks for air quality monitoring, and robot design and locomotion strategies.

Wednesday, March 11 9:00 - 10:40

WA1: Robotics, Automation and Data Fusion

Room A
Chairs: Alain Pegatoquet, Hans-Peter Schmidt
9:00 Context-Aware Sensor Adaption of a Radar and Time-of-Flight Based Perception Platform
In this paper we present an approach to enhance the perception quality of a multi-sensor system by considering the context information. The sensors' configuration and processing parameters are automatically adjusted based on the system's current context. The platform's context state is obtained by combining data from the perception sensors and from additional context sensors. This context information is utilized to react to environmental influences on the perception sensors as well as to degrade affected sensors before fusion. We demonstrate the presented approach with a multi-sensor platform equipped with Time-of-Flight (ToF) cameras, radar sensors and multiple context sensors. The Robot Operating System (ROS) architecture of the platform is extended to support the automatic adaption of sensor parameters. We exposed the platform to different scenarios in order to show the context-aware adaption of the sensor configuration as well as the automatic sensor degradation. Our context-aware self-adaption approach leads to a significant performance improvement compared to a static system.
Presenter bio: Josef Steinbaeck received his Master's degree in Information and Computer Engineering from the Graz University of Technology, Austria in 2016. He is currently a PhD student at Infineon Technologies in Graz with the focus on automotive/robotic environmental perception sensors.
9:20 A Comprehensive Study of Performance-Autonomy Tradeoff on Smart Connected Glasses
Embedded detection systems presented in the form of wearable devices become very useful in the health care area. Due to limitations as size, energy and computational resources, the power consumption must be studied. Considering performance constraints during the study of power on smart connected glasses allows to propose reliability and good autonomy. An optimization methodology based on both a system level power modeling approach and a performance characterization of the detection system through several performance metrics is introduced in the present article. A comprehensive study for finding the system configuration that offers the best tradeoff between performance and autonomy, is proposed. Both accuracy and the relationship between sensitivity and specificity provide similar results about the performance of the blinks detection.
Presenter bio: PhD Candidate at the University Cote d'Azur, working on the optimization of the energy consumption on the Ellcie-Healthy Smart Connected Glasses, through system-level modeling approaches under QoS constraints. This project is a collaboration between the Ellcie-Healthy start-up and the LEAT laboratory.
9:40 Robust hand tracking method by synchronized high-speed cameras with orthogonal geometry
In this paper, we suggest an intuitive hand motion tracking method using synchronized high-speed cameras with orthogonal geometry. Three synchronized high-speed cameras that capture and process images at the speed of 1,000 frames-per-second concurrently are intentionally arranged so that any pair of optical axes of the cameras cross each other orthogonally. The combination of the high speed with the orthogonal arrangement contributes to robust tracking during dynamic motion with the high-speed three-dimensional reconstruction under occlusion. We present a preliminary hand tracking system using three cameras adopting a minimalist design and also study its feasibility and effectiveness both mathematically and experimentally.
Presenter bio: Hyuno Kim is a project assistant professor at Department of Information Physics and Computing, The University of Tokyo in Japan. His current research interests include high-speed camera networks, high-speed visual feedback control, and robotic manipulation using high-speed machine vision.
10:00 Sensor Fusion for Analysis of Gait under Cognitive Load: Deep Learning Approach
Human mobility requires substantial cognitive resources, thus elevated complexity in the navigated environment instigates gait deterioration due to naturally limited cognitive load capacity. This work uses deep learning methods for 116 sensors fusion to study the effects of cognitive load on human gait of healthy subjects. We demonstrate classifications, achieving 86% precision with Convolutional Neural Networks (CNN), of normal gait as well as 15 subjects' gait under two types of cognitive demanding tasks. Floor sensors capturing multiples of up to 4 uninterrupted steps were utilized to harvest the raw gait spatiotemporal signals, based on the ground reaction force (GRF). A Layer-Wise Relevance Propagation (LRP) technique is proposed to interpret the CNN prediction in terms of relevance to standard events in the gait cycle. LRP projects the model predictions back to the input gait spatiotemporal signal, to generate a "heat map" over the original training set, or an unknown sample classified by the model. This allows valuable insight into which parts of the gait spatiotemporal signal have the heaviest influence on the gait classification and consequently, which gate events are mostly affected by cognitive load.
10:20 Measurement Platform for Physical-Layer Analysis of Industrial and Automotive Ethernet
Robustness of communication is essential for many automotive and industrial Ethernet applications due to real-time and safety features that are required. Existing data transmission structures have to be used for higher transmission rates in some application. Likewise, some new installation require cost-effective cabling with non-ideal transmission characteristics such as unshielded single twisted pair. This enables economic feasibility of Industrial Internet of Things networking of simple sensors and actors. We introduce a universal platform for physical layer analysis for such physical layers. It covers communication channels up to some 10 Mbit/s which are typically found in industrial and automotive Ethernet. We present a combined soft- and hardware solution. It comprises an end to end communication and facilities spectrum, correlation and constellation analysis. Application to and results for unshielded twisted pair cabling are presented.
Presenter bio: Master and Phd Electircal Engeinerng and Information Technology from TU Munich Faculty Postion, TU Munich Industry, SiemnesBerlin Germnay 1997 Professor OTH University of Applied Sciences additioanl 2008 Head aia Automation Institute 2014 Head AUT + aia joined workgroup

WB1: Sensors for Smart Mobility

Room B
Chair: Eric Matson
9:00 Initial Evaluation of Vehicle Type Identification using Roadside Stereo Microphones
A key feature of Intelligent Transport Systems (ITS) is the ability to detect and identify vehicles. In this paper, we put forward a stereo microphone-based system capable of detecting and identifying the type of individually, sequentially, and simultaneously passing vehicles in multi-lane environments based on their sound. We find that our proposed system shows improved performance over single-microphone systems thanks to its improved sequential and successive vehicle detection performance. Initial evaluation results using sound data collected from roads on a university campus show a classification accuracy of 95.01%.
Presenter bio: Billy Dawton is a doctoral student with the department of Information Science and Electrical Engineering at Kyushu University, Japan. His current research focuses on audio-based ITS applications, in particular low-cost, low-complexity acoustic vehicle detection and identification systems.
9:20 Location Sensing using QR codes via 2D camera for Automated Guided Vehicles
Automated Guided Vehicle (AGV) have gradually played a key role in many industrial systems such as factories and logistics. The key factor for AGV to be adopted is cost effective and accurate location sensing capabilities. In this paper, QR codes are used as artificial landmarks for AGVs to perform location sensing. This work differs from conventional floor-grid-based QR codes for AGVs as location tags, this work proposes to use the distance and angle between a camera attached on an AGV and the QR codes that are strategically placed around a facilities to compute the exact location of the AGV. Experiment is conducted to test on the accuracy of the computed distance and angle. The proposed localization approach allows self-localization on multi-AGV systems and by using QR codes, the ambiguity of finding unique landmarks in a facility can be eliminated.
9:40 Coarse Object Tracking Technique for Point Clouds
Object detection and avoidance are the core functions performed by autonomous vehicles. To this end, object tracking over time and speed calculation are of the utmost importance in object avoidance algorithms. Autonomous vehicles need to be able to reliably and quickly estimate the speed of an object and track it over multiple frames. Object prioritization based on speed is of the utmost importance in collision avoidance. This is of greater interest in connected vehicles and vehicular swarms where messages indicating objects that are headed for an accident can be sent to other vehicles so they can avoid a collision. This paper proposes a fast and accurate method of tracking distances moved by objects between frames.
Presenter bio: Academics: M.Tech and PhD from IIT Madras in Communication Networks Areas of Research: Wireless sensor networks, Cyber Physical Systems, Optical Networks, Embedded Systems
10:00 Accurate Perception for Autonomous Driving: Application of Kalman Filter for Sensor Fusion
Object tracking is a foundation task for autonomous driving. An increasing amount of work applies sensor fusion to facilitate the tracking results. However, sensor fusion alone still cannot reach a desirable accuracy in real road conditions, because of the sensor noises and complexity of the motion dynamics. Bearing this in mind, we propose a method that applies Kalman filter on the LiDAR and radar sensor fusion to improve the accuracy in object tracking for the autonomous driving system. We evaluate our approach on the Udacity dataset. Results demonstrate that the Kalman filter drastically improves the final measurement compared to using sensor measurement alone. The work verifies the effectiveness of employment Kalman filter to facilitate the performance of sensor fusion measurements for the autonomous driving system.
10:20 Enabling Live State-of-Health Monitoring for a Safety-Critical Automotive LiDAR System
In the next few years, modern vehicles will integrate the next level of Advanced Driver-Assistance Systems (ADAS) such as Light Detection and Ranging (LiDAR) which will be one of the key enabler for autonomous driving. Autonomous driving will be in charge for controlling the vehicle without any inputs of a passenger. This requires highly robust and reliable components and systems. In general, mechanical defects are detectable through vibrations or noise changes but for semiconductor components these capabilities are not available. Semiconductor components fail silently and abrupt without any prior information and this could lead to fatal accidents when systems fail during autonomous driving phases. In this publication, we are introducing a novel state-of-health monitoring system for automotive LiDAR system that is capable to economically record the component history and automatically processes these data to the statistical Failure-In-Time (FIT) Rate that is primarily used in the Automotive domain such as in the ``ISO 26262 - Road Vehicle Safety'' standard.
Presenter bio: I studied Computer Engineering at Graz University of Technology and right now i am a PHD student at the same university. My research interests are in safety-critical embedded systems and functional safety in general.

Wednesday, March 11 11:10 - 12:30

WA2: Nonlinear Sensing: From Materials to Systems

Special Session
Room B
Chair: Behraad Bahreyni
11:10 A Multi-Jerk Equation Emulator Circuit Exhibiting Various Chaotic Behaviours
This paper presents the design, mathematical background and performance studies of a new chaotic emulator circuit. The circuit exhibits various chaotic phenomena by emulating multiple jerk equations. The circuit is designed using simple and easy-to-use components like operational amplifiers, passive components, and a single switch. It uses an inductor-free and diode-less architecture; thus eliminating the nonideal parameters of these elements. The circuit operates in three modes; each mode exhibiting a different chaotic behaviour. Detailed working of the circuit and its ability to follow jerk equations are explained in the paper. This is followed by performance studies, first using a SPICE software and then using experimentation. These studies validate the capacity of the proposed circuit to emulate multiple jerk equations and various chaotic attractors. The experimental studies corroborate well with simulation and results mentioned in previous works.
Presenter bio: Denil V Robinson is pursuing Btech in Electronics and Communication Engineering from Indian Institute of Space Science and Technology, Thiruvananthapuram. His research interests are in the area of Instrumentation and Measurement.
Denil V Robinson
11:30 A Digital Signal-Conditioner for Resistive Sensors and its Utility for Linearizing GMR-based Magnetometer
In this paper, a novel relaxation-oscillator based resistance to digital converter is proposed. The converter processes the resistive sensors with a circuit employing a single reference voltage and possesses low execution time. The circuit is simple and it provides many advantages like independence from amplifier gain, switch resistance, etc. The performance of the circuit is simulated and emulated for different types of resistive sensor configurations, and results are presented in this paper. The maximum nonlinearity observed from the above studies is 0.2 %. The possible error sources are brought out to show the accuracy of the proposed work. Finally, the proposed circuit is interfaced with the commercially available GMR sensor to develop a GMR-based magnetometer. Results show that the proposed circuit provides better performance than the existing works.
Presenter bio: Research scholar
11:50 Feasibility study of fiber taper acoustic sensor by utilizing time domain reconstruction
A fiber taper acoustic sensor utilize the ability of taper structure to modulate vibration signal into the fiber transmission power. Currently, there is a lack of detailed study on the bending effect of fiber taper to the fiber transmission power. In this paper, wide-angle propagation method is used to simulate the effect of fiber taper bending. The reconstructed time-domain signal and time-domain signal modulated from vibration has a nonlinear negative relationship and has a correlation coefficient value of 0.9979
12:10 Demonstration of a Nonlinear Angular Rate Sensor based on Internal Resonance
This paper reports on the design, fabrication, and characterization of an H-shaped tuning fork microresonator with 2:1 internal resonance as an actuation mechanism. The nonlinear principle of operation addresses major challenges in MEMS Coriolis vibratory gyroscopes: eliminating the mode-matching, minimizing instability and drift due to mechanical cross-coupling, and generating a wide operating frequency range with high-signal gain and less sensitivity to fluctuations in driving frequency. The rate measurement relies on capturing the half-order subharmonic response of the device while undergoing the angular velocity. The micromachined resonator is fabricated by Teledyne DALSA Inc. The experimental finding demonstrated the prominent M-shaped nonlinear resonant curves due to a frequency ratio close to 2:1. The microresonator is nominally operated in the overlap region between the forward and backward frequency sweeps, where the signal gain is less sensitive to frequency fluctuations. Experimental rate characterization of the microresonator revealed a linear dynamic range of 220 deg sec-1 with a sensitivity of 0.011 mV deg-1 sec-1 using an 80V DC polarization voltage. The experimental results of the microresonator showed the induced oscillations in the so-called pendulum mode by Coriolis force coupling, despite a clear disparity on natural frequencies of the desired modes.
12:30 A Multi-Level Information Fusion-based Deep Learning Method for Vision-based Defect Recognition
Vision-based defect recognition is an important technology to guarantee quality in modern manufacturing systems. And deep learning (DL) becomes a research hotspot in vision-based defect recognition due to the outstanding performances. However, most of the DL methods require a large sample to learn the defect information. While in some real-world cases, it is difficult and costly for data collecting, and only a small sample is available. Generally, a small sample contains less information, which may mislead the DL models, so that they cannot work as expected. Therefore, this requirement impedes the wide applications of DL in vision-based defect recognition. To overcome this problem, this paper proposes a multi-level information fusion-based DL method for vision-based defect recognition. In the proposed method, a three-level Gaussian pyramid is introduced to generate multi-level information of the defect, so that more information is available for model training. After the Gaussian pyramid, three VGG16 networks are built to learn the information and the outputs are fused for the final recognition result. The experimental results show that the proposed method can extract more useful information and achieve better performances on small-sample tasks, compared with the conventional DL methods and defect recognition methods. Furthermore, the analysis results of the robustness and response time also indicate that the proposed method is robust for the noise input, and it is fast for defect recognition, which takes 13.74ms to handle a defect image. DOI:10.1109/TIM.2019.2947800