Program for 2019 Zooming Innovation in Consumer Technologies Conference (ZINC)
Wednesday, May 29 8:00 - 17:00
Welcome and Registration
Room: Hall Middle
Chair: Jelena Gajic
Wednesday, May 29 9:00 - 10:15
Galvanizing Consumer Technology, Image and Video Processing
Chairs: Mališa Marijan, Mario Vranjes
- 9:00 Drone Aided Inspection during Bridge Construction
- This paper presents a case of filming a bridge in Novi Sad during its construction for the purpose of inspection of construction works carried out by the Spanish - Italian consortium AZVI. Twice a month the photographs and high-resolution films shot by a drone were sent to AZVI. In this way, the contractors had up-to-date information on the progress of the works without the need for a permanent presence of the company's engineers. Drone shooting has enabled high-resolution images and movies which allow quality remote inspection, accompanied, naturally, with occasional arrivals at the site.
- 9:15 Field Trial Results for Portable Indoor Reception of the DVB-T2 Base Profile
- This paper describes the field trials carried out in the Bucharest region in Romania for testing and analyzing the performance of the second generation of the Digital Video Broadcasting Terrestrial standard using the base profile for portable indoor reception. Test results are presented for the terrestrial broadcasted public programs of the Romanian National Television. The results consist of the values of the Carrier to Noise ratio, Modulation Error Rate and Link Margin in different geographical locations and receiving conditions. These values can update those provided by previous studies and provide planning tools for the Digital Terrestrial Television implementation using the second generation of the Digital Video Broadcasting Terrestrial standard. The results proved that the portable indoor reception in the Bucharest region for the Romanian National Television programs is possible, and this may be used to improve the end user experience with an additional television service to the existing cable or satellite television, since the digital terrestrial television has the advantage of mobility reception.
- 9:30 Psychophysiological Stress Control via Heart Rate Entrainment
- The aim of this work is to develop a real-time system able to classify and passively reduce stress. Heart Rate (HR) and Skin Conductance Response (SCR) are used for stress classification. Entrainment phenomenon is exploited as a possible way to reduce stress via an auditory stimulus. Support Vector Machine (SVM) produces 84\% of stress classification accuracy and slight effect is obtained for stress reduction.
- 9:45 Analysis of Still Image IP Transmission in the Tactical Communication Network
- This paper presents analysis of still image IP transmission in the tactical network. The analysis is provided through the delivery time, packets sent/received/lost and the display of the images after the reconstruction on the receiving side. Image transmission is analyzed through multiple scenarios - without and with compression, without and in the presence of jamming. It has been shown that delivery time can be significantly reduced by data compression, without significant loss of visual information. It has also been shown that for analyzing the quality of transmission and usefulness of received information, it is better to use the display of reconstructed images than the degree of packet losses. Furthermore, it has been shown that in the tactical networks of limited bandwidth, it is very important to comply bit rates from source of information to the end user, and in this way to avoid buffer memory overflow and packet losses.
- 10:00 Biometric Security Standardization
- By developing and increasing the application and use of Information Technology (IT), both hardware and software, the way of life and habits of people and societies using these technologies changes. In the face of this IT development and changes in society that you have with you, standards that regulate the use and development of such technologies must be created and changed. In everyday life, we are asked to identify with PINs, passwords, personal documents, and more recently with biometric data. This paper provides an overview of organizations involved in biometric standardization and the biometrics standards that have been developed and are currently in progress.
Student session: Galvanizing Automotive solutions 1
Chairs: Ratko Grbić, Bogdan Pavkovic
- 9:00 Surround view algorithm for parking assist system
- Modern vehicles use different advanced driver-assistance systems (ADAS) in order to make driving safer and more comfortable. One of them is a system that uses a set of in-vehicle cameras and provides the driver a top-view image of the space around the vehicle, thus helping the driver in parking. The main part of that system is the algorithm which processes the frames acquired simultaneously from four in-vehicle cameras located at different sides of the vehicle and creates the final top-view image of the space around the vehicle. In this paper an new algorithm for that purpose is designed, which consists of two main parts, one for performing camera calibration, the other for generating the top-view image. The implemented camera calibration is based on calibration patterns and calculates camera parameters which are then used to eliminate image distortion due to usage of fish-eye cameras. For generating the top-view image of the space around the vehicle, different geometric operations including distortion correction, perspective transformation, and image stitching, are applied to the input frames acquired by different cameras. Algorithm performance are tested using four real automotive fish-eye cameras fixed to the vehicle model and connected to the ADAS development board.
- 9:15 Object detection and object tracking in front of the vehicle using front view camera
- Modern vehicles are equipped with the different systems that help driver in the driving process ensuring safer and more comfortable driving. These systems are called Advanced Driver Assistance Systems (ADAS) and are step toward fully autonomous driving. The integral part of autonomous driving is an object detection and tracking by using front view camera which provides necessary information for emergency braking, collision avoidance, path planning, etc. In this paper, one possible approach to object detection and tracking in autonomous driving is presented. Two object detection methods are implemented and tested: Viola-Jones algorithm and YOLOv3. The Viola-Jones algorithm is used to create object detectors which detections are tracked in a video sequence. Nine object detectors were trained and they are divided into four groups (vehicle detectors, pedestrian detector, traffic light detector and traffic sign detectors). In second case, the YOLOv3 model was used for object detection. Both methods are evaluated in terms of accuracy and processing speed. For the purpose of object tracking, Median Flow tracking method and correlation tracking method are implemented and evaluated.
- 9:30 Driver monitoring using the in-vehicle camera
- One of the possible ways to reduce the number of accidents in traffic is to track the driver's condition. In this paper, a new driver monitoring algorithm, which operates on a video captured from the camera located in the instrument panel in the vehicle, is presented. The algorithm examines the presence of the driver's head in each frame and checks whether the driver has unusually long-closed eyes, whether the driver looks sideways, and whether the driver's head has dropped for a long time. The algorithm is tested using ten different videos recorded under different light conditions. The algorithm generally achieves high performance, while for some lighting conditions a certain algorithm upgrading would be necessary to maintain a high level of accuracy.
- 9:45 Software solution stack for data transfer on a frame grabber platform
- With the progress of automotive industry arose a need for development of fully autonomous vehicles. Such vehicle is equipped with sensors and cameras designed for monitoring its surroundings. Data received from these sensors is then processed in the vehicle's embedded system which contains algorithms that help driver in a driving process and are a step towards a fully autonomous vehicle. Due to the problem complexity, these algorithms are usually based on machine learning methods and thus inherently require large amounts of data for successful training. Data used for training is mostly video content of situations from actual traffic. To record that video content specialized hardware and its accompanying software is needed. This paper focuses on developing software for existing AMV Grabber hardware board. Whole stack of software was developed, from low-level application which controls the hardware directly, through device driver which enables communication between the board and PC, to PC application which enables users to control the hardware indirectly, i.e. to send commands to the board to start or stop recording, and which is used to receive data from the board and store it on non-volatile memory of the PC. Finally, measurements were done to display overall system performance.
- 10:00 Trajectory Planning for Autonomous Vehicle Using Digital Map
- In this paper, an algorithm for autonomous vehicle trajectory planning is proposed. The proposed algorithm consists of two parts. In the first part, transformation of given coordinates from Cartesian system to Frenet-Serret system is applied, along with interpolation of points in trajectory. In the second part, most efficient trajectory is being chosen and optimized. The algorithm was developed in Udacity's simulator, where it was tested and compared with other similar algorithms. The algorithm shows promising performance through testing.
Wednesday, May 29 9:00 - 17:00
Chair: Milos Stankovic
Wednesday, May 29 10:15 - 11:00
Keynote1: "How to innovate in AI era?"
How to innovate in AI eraRoom: A1
Chairs: Gabriele Foresi, Velibor Ilic
- 10:15 How to innovate in AI era
- Artificial intelligence promises both to improve existing goods and services, and, by enabling the automation of many tasks, to greatly increase the efficiency with which they are produced. But it may have an even larger impact on the economy by serving as a new general-purpose "method of invention" that can reshape the nature of the innovation process and the organization of R&D. This presentation considers this possibility in three interrelated elements. 1. History of AI and world current stage among China, USA and Europe 2. Patent research and methods how to address challenges 3. Deployment techniques that speed up the story further as Machine Learning Canvas
Wednesday, May 29 11:00 - 11:30
Galvanize your network: Coffee break
Rooms: A1, B1, C1, Caffeteria
Chair: Jelena Gajic
Wednesday, May 29 11:30 - 12:15
Keynote2: "Standards and conformity assessment relevance for Consumer Electronics"
Standards and conformity assessment relevance for Consumer ElectronicsRoom: A1
Chairs: Dragan Kukolj, Branislav M. Todorović
- 11:30 Standards and conformity assessment relevance for Consumer Electronics
- Presentation will interpret international, European and Serbian standardization issues relevant for current public and business context. Focus will be on ICT as ongoing industrial revolution the civilization has been facing, and ISO, IEC and ITU as apex international standards organizations. Standardization relevance for UN, WTO and EU will be also mentioned. Understanding of key points of standardization as leverage for any undertaking may additionally rationalize research and development efforts. Presentation will treat both standard users and standards developers. Practical examples will in short cover typical standard structure for proper understanding, and CE most relevant technical and other bodies and International Standards. Management system certification importance for consumer confidence also deserves participant's attention.
Wednesday, May 29 12:15 - 13:00
Opening: Let the ZINC begin
Chair: Gordana Velikić
Wednesday, May 29 13:00 - 14:15
Galvanize your network: Lunch
Rooms: A1, B1, C1, Caffeteria
Chair: Jelena Gajic
Wednesday, May 29 14:15 - 15:30
Open Day (Serbian)
Chairs: Ivan Kastelan, Lea Vujovic
Galvanizing Software and Computer Engineering
Chairs: Marijan Herceg, Mario Vranjes
- 14:15 The VR Simulation Of Hidrological Data
- Urban plan "Extension of the Zemun quay" was made from the need for access to the population of the Danube coast. The plan envisages the construction of a promenade and facilities of various contents in a zone of quay which is not defended from high water level. The visualization of the plan was made for the optimal level of the Danube River water level in immersive virtual reality (VR). As a variety of construction is planned at this location, simulation of the extreme hydrological data of the Danube River was also performed. The aim of visualization and simulation of this specific location and the impact of natural factors on the objects is to encourage and assist decision-makers in building investors. The direct experience gained in VR will enable investors to determine which type of facility they will invest in. This paper will show basic research activities performed for the purpose of simulating the consequences of the extremely high water levels of the Danube.
- 14:33 Educational microprocessor implementation on ZYNQ reconfigurable system on chip
- In this paper we present the design and implementation of an educational microprocessor named ZEdulent. It emerged as evolution of Edulent processor. ZEdulent is designed to be implemented on ZYNQ reconfigurable system on chip. Accordingly, it is tested on Digilent ZYBO development board. Software support running on Linux is designed in Qt framework with graphical user interface focusing educational goals. Hardware and software implementation details are given and compared to previous implementations.
- 14:52 Implementation and signal processing of colorimetric probe
- The idea of this paper is to implement a colorimetric probe that will be used for measuring color of printed color samples obtained using digital printing. The method will be based on the estimation of the spectrum of electromagnetic radiation (in the visible range) over previously known radiation values at certain wavelengths. The processing and filtering of signals from the colorimetric probe are explained in detail. By selecting the appropriate parameters and methods a fast response and corresponding stability of the system is achieved.
- 15:11 One method for hybrid playback and recording of digital TV channels on embedded devices
- Todays, most embedded TV devices on field are powered by Linux operating system. They possess different front ends for accusation of TV signal: terrestrial, satellite, cable or over internet. Having multiple front ends, results are hybrid devices which can present user TV content from multiple sources. Multiple sources provide richer content, better experience and more stable delivery, but also introduce duplicate TV services in channel list. This paper will analyze the above-mentioned problem and present solution for providing unique channel list to user with best available content quality. Paper will also cover recording use cases and resource conflicts.
Wednesday, May 29 15:30 - 17:00
BENEFIT Seminar 1
Central Audio Library at the University of Novi Sad as a New Teaching and Learning MethodologyRoom: A1
Chair: Velibor Mihic
- 15:30 Central Audio Library at the University of Novi Sad as a New Teaching and Learning Methodology
- Central Audio-Library of the University of Novi Sad (Serb. Centralna audio-biblioteka Univerziteta u Novom Sadu - CABUNS) will be presented to both kind of users - teachers and students. CABUNS automatically produces audio editions of textbooks and PPT presentations using Text-to-Speech (TTS) technology for conversion of text to synthesized speech. It is organised as a web service. Authorised teachers just submit textbooks and presentations to CABUNS server and TTS engine creates their audio editions. Afterwards audio materials are available to all authorised students using mobile phones, tablet, laptop or desktop computers, or TVs - practically any device connected to Internet. All seminar attendants will have oportunity to try the system, hear and evaluate the synthesized speech. SWOT analysis will present advantages and limitations of CABUNS as a new teaching and learning methodology based on TTS and ICT. CABUNS can initiate a network of audio libraries in the region. The architecture of CABUNS, as well as software tools and hardware resources used in the project, will be presented in the second part of seminar. The third part of seminar is dedicated to the presentation of TTS development based on deep learning paradigm, challenges and achievements for Serbian TTS, including perspectives of voice and style conversion - which could enable students to listen lectures synthesized in voice similar to their professor's voice. Finally, a mini debate will be initiated, related to open questions and non solved challenges, as well as to perspectives of audio libraries based on TTS as new methodology for teaching/learning at different level of education.
Interviews: sparks of creativity
Chair: Konstantin Glasman
Wednesday, May 29 15:30 - 16:30
Chair: Milan Z. Bjelica
- 15:30 Performance analysis on TCP/IP Audio Streaming in Point-to-Point communication
- UDP has always been a choice number one when it comes to the multimedia streaming systems. Not solely because TCP congestion control does not ensure delivery in a timely fashion, but also because these systems can tolerate data loss. Nonetheless, verification systems, including ones used in DSP audio field, which require both timely and reliable delivery are bringing much more constraints to the mix. The advantage of these systems, on the other hand, is that they are used in the simplest, point-to-point communication topology. In this kind of topology and with latest advances in both Ethernet hardware and embedded software, this paper examines the idea of using TCP instead of UDP application protocol for streaming high-quality audio between two points - device designed for audio system verification, and a personal computer, engineers working station. The main idea of this paper is reducing engineering efforts and development costs by consuming less time for making custom UDP protocol which would satisfy both full-duplex high throughput and reliable data exchange requirements.
- 15:35 Electric motors of large consumer products, challenges and trends from the perspective of power efficiency improvement through modern cutting technologies
- Electrical motors are indispensable components in numerous large consumer goods, most notably home appliances, besides electronic circuitry and mechanical parts. Their large extent in use rises several important challenges to be addressed in a fiercely competitive globalized marked, like production costs vs. intrinsic product quality and performance, including here energy efficiency, as one of the most important factors. Several reasons to be reviewed in our present work lead to adopted standards and regulation to provide an effective and proactive answer to the problem. Then, the study focuses on reviewing one of the main sources of overall efficiency degradation, namely energy steel losses, analyzed through specific methods and illustrated with comprehensive experimental results. Apparently insignificant, small energy consumption reduction may globally provide an important impact on large consumer goods, manufactured on an industrial scale. Finally, conclusions are drawn to envisage further trends and material investigative methods for electrical machines manufacturing, especially from the perspective of generalization of IE4 efficiency standard.
- 15:40 One solution for Over The Air software updates in the vehicle on the Adaptive AUTOSAR platform
- Automotive software in modern vehicles is becoming very complex and the main problem for manufacturers is to ensure that new features, bug fixes and improvements are quickly delivered to vehicles. Over-the-Air (OTA) updates are considered as a faster way of suppling new software while reducing the cost as well. This requires the development of a platform with possibility of dynamic deployment and update of applications which is Adaptive AUTOSAR. Thisplatform shall not violate proper work of safety critical ECUs, and shallkeep the system safe and stable at all times. In this paper, a solution for vehicle update which includes IoT techonologies, Adaptive AUTOSAR extensions and installation flow as well,is presented.
- 15:45 Driver Monitoring implementation in Adaptive AUTOSAR environment
- Advanced Driver Assistance System (ADAS) algorithms are significant in the modern automotive industry. The algorithms have changed the approach to traffic efficiency, and safety, two important issues of the industry. ADAS algorithms have real time execution and low memory consumption requirements. Therefore one of the main challenges is to satisfy these requirements without losing the reliability. One of the most common components in driver assistance systems is drowsiness detection, because it is one of top causes of traffic accidents. In this paper, we present driver drowsiness detection solution which is implemented on the Adaptive AUTOSAR platform.
- 15:50 The detection of traffic participants using YOLO algorithm
- With the outstanding technological advancement, new areas of research such as autonomous driving, are born. Artificial neural networks are used in this area, because of the large amounts of data and the complexity of the problem. Classical approaches to problem solving traffic participants detection, as a component needed for autonomous driving, have the problem of the slow response time, therefore they can't be used in real-time systems. Modern architectures of artificial neural networks and modern algorithms, such as YOLO algorithm solve the problem of the slow response time without precision losses. This paper provides the demonstration of the usage of YOLO algorithm for the traffic participants detection.
- 15:55 Implementation of incremental linking for MIPS64 architecture in GOLD linker
- Incremental linking is a mechanism which lowers the program's link time. This means that all program components will not be linked every time, just the ones that have changed since the last linking. This mechanism is used exclusively during the development of the program. The task of this thesis is to implement the incremental linking mechanism for MIPS64 computer architecture in GOLD linker. The solution is verified using built-in incremental linking tests in GOLD linker.
- 16:00 Gathering diagnostic data over one UTP cable ethernet communication interface - BroadR-Reach
- Automotive industry is undergoing changes nowadays in terms of Ethernet due to cost benefits with high bandwidth, easy integration and light cable weight. Although Ethernet has been used in LAN networks for decades it is a new technology for automotive. Automotive Ethernet, based on BroadR-Reach standard, has now started to expand into areas that had been dominated by CAN, LIN, FlexRay and Most and because of new transceiver circuit, some test methods need to be modified. In this paper, we test the physical communication mechanism for BroadR-Reach, and display data in application. Some test cases are applied to TJA1100 from NXP. Although these test results are related to chip manufacturers, the analyisis results still have guiding sigificance for the applicaton of Ethernet to avoid dangerous conditions.
- 16:05 Emulation of I2C interface system via serial connection
- Topic of this thesis is emulation of I2C interface system via serial communication. It's necessary to connect development board with personal computer via serial connection, enable sending and receiving I2C telegrams via serial connection. The last telegram is printed on LCD display, where is necessary to show device address, register address, type of access (read or write) and data value. Inside development board it was necessary to implement attendant device with sixteen 16-bits registers, all of registers have ability of reading and writing via I2C telegrams.
- 16:10 One solution of vehicle control software based on camera in ROS environment
- Software in automotive industry today is a very popular branch and brings many new challenges to the world of engineering. Recently, autonomous driving became one of the biggest challenges of this field. Goal is to develop a system capable of controlling vehicle almost completely independently. Work presented in this paper defines several modern autonomous driving algorithms and simulate them on modern platform. These algorithms are working with sensors attached to vehicle, such as camera, radar, lidar, etc. and include functionalities such as keep lane, controlling vehicle when "Stop" sign is detected and adapting vehicle speed when speed limit sign occurs in front of camera sensor.
- 16:15 Realisation of an object audio rendering module on an SoC architecture based on ARM Cortex-A family of cores
- In the last decade, there have been major changes in both quality and quantity of audio and video content consumption. One of the key features of newer technology is an accurate spatial reproduction of the sound signal, regardless of the speaker configuration. This was made possible by the 3D object concept. Using rendering, objects are dynamically allocated to the speakers present in the system configuration. This paper describes an implementation of object audio rendering module on such architecture based on ARM Cortex-a family of cores.
- 16:20 Android widget applications development
- One implementation of the client Android widget application is proposed. The implemented application is used for tracking states of controllers inside the zones of a smart home. The application is implemented using Android Studio development environment.
- 16:25 Google Assistant integration in Live Channels application for Adroid OS
- The majority of existing speech recognition tools provide the result of the speech processing as a free form textual output or structured form textual output. In order to use obtained outputs in an existing TV applications, it is necessary to detect patterns that correspond to commands such as channel up/down, volume up/down, mute/unmute, etc. This paper presents one solution for using TV application with the Google Assistant support for Android OS.
Wednesday, May 29 17:00 - 17:30
Galvanize your network: Coffee break
Rooms: A1, B1, C1, Caffeteria
Chair: Jelena Gajic
Wednesday, May 29 17:30 - 17:45
Keynote3 (Panel Intro): "Do you speak the language?"
Do you speak the language?Room: A1
Chairs: Nemanja Kovacev, Gordana Velikić
- 17:30 Do you speak the language?
- Do you speak the language?
Wednesday, May 29 17:30 - 19:00
BENEFIT Seminar 2
Automotive Communication (Car-2-X), Digital Image and Video Processing in Automotive SystemsRoom: B1
Chair: Marija Antic
- 17:30 Automotive Communication (Car-2-X), Digital Image and Video Processing in Automotive Systems
- Focus of the lecture is to bring the audience closer to the hot topic of automotive engineering with special focus on 2 prominent fields - Computer Networks and Image Processing. In the Computer Networks part we will focus on exploring the trends of Car-2-X communication: communication that can take place inside (wired) and outside (wireless) of the vehicle. We will explore the specific automotive communication protocols like LIN, CAN, FlexRay, and BroadR Reach and specific challenges that automotive environment should overcome. Furthermore, we will highlight applications of wireless technologies to enable safer roads and driving. In the image processing part we will explain why applications based on image processing and computer vision are crucial for autonomous vehicles. Additionally, the focus will be on the analysis of complex Advanced Driver Assistance Systems (ADAS) algorithms and discussion about image processing techniques used for their design.
Wednesday, May 29 17:45 - 18:00
Keynote4 (Panel Intro): "Will the laser kill the back pain?"
Will the laser kill the back pain?Room: A1
Chairs: Milenko Beric, Nahum Gershon
- 17:45 Will the laser kill the back pain?
- Will the laser kill the back pain?
Wednesday, May 29 18:00 - 19:00
Panel: "Consumer Technology Meets Health Care"
Consumer Technology Meets Health CareRoom: A1
Chair: Nahum Gershon
- 18:00 Consumer Technology Meets Health Care
- Panelists: Nahum Gerhson - moderator, Milenko Beric, Gordana Velikic, Nemanja Kovacev, Lucio Cabattoni, Gabriele Foresi, Predrag Filipov, Danica Sazdanic-Velikic, Audience Living in a remote place, Mike felt really bad this evening and he wished that he could see a doctor. But, it was not easy and practical for him to drive 3 hours to the nearest hospital. Luckily, he had a new device with which he could perform some basic measurements like heart rate, temperature, ear and throat check, and temperature. A physician at the hospital was seen on the screen of the small device and after looking at the results of the tests and talking with Mike, he sent recommendations for the patient. In 2 hours, Amazon delivered the prescribed medication at Mike's doorsteps. This illustration is not science fiction anymore. Consumer technology has gone a long way since the days when devices were produced unchecked and were of questionable quality. There are now, for example, small and portable devices to measure blood pressure with a medical accuracy level that also measure the number of steps, sleep quality and reports on incoming calls. Such devices are expected to revolutionize health care as we know it through mobile health. Not just enabling patients in remote places to get the health care they need, but also enabling everybody to get some treatments at home without visiting the physician's office. This could allow the early release of specific patients from the hospital when the treatment and rehabilitation could be done at home under the remote care of the physician or practitioner while the data is being captured and included in the patient's medical records. This may also revolutionize the existing overall model of the interaction of patients with their physicians and other health care practitioners. There are some challenges, of course. For example, not all humans are alike and we would need to evaluate (e.g., on a psychological level) who from the patients will not respond well to this remote or often remote health care as compared to face-to-face treatment by the physician. Other challenges include understanding how to relate to a collection of sensors (as compared with one sensor) and relating to the body as a system and not just a point in space as some of the current devices do. These developments could dramatically improve the delivery of services to the general public and to specific groups of patients and increase the magnitude and depth of the collaboration between the patients and their medical providers. This is an opportunity for the consumer technology community to increase the breadth and depth of our impact on health.
Wednesday, May 29 20:00 - 23:59
Rooms: A1, B1, C1
Chairs: Gordana Velikić, Milan Vidakovic
Thursday, May 30 8:00 - 9:00
Welcome and Registration
Rooms: A1, B1, C1
Chair: Jelena Gajic
Thursday, May 30 9:00 - 10:15
Galvanizing IoT, Smart Energy Management
Chairs: Istvan Papp, Boris Radin
- 9:00 Cloud Based Data Acquisition via IoT for Electric Power Quality Monitoring
- This paper presents the usage of Cloud and IoT technology in data acquisitions for power quality analysis. Cloud technology allows measuring devices to work using IoT infrastructure, where devices over the Cloud are networked on the Internet and can quickly and efficiently transfer information about the state of power quality in power network. The initial information provided by the measuring devices is the measured active and reactive power, as well as the magnetizing current harmonic content. The collected data are stored in the Cloud, where later accurate and transparent data analysis is possible with the aim of estimating the DC voltage offset in power network, which most commonly appears as the result of a photovoltaic power grid on the network.
- 9:18 Design and development of a smart attendance management system with Bluetooth low energy beacons
- The tracking of attendance is a ubiquitous task in many institutions and during organised events. This, along with subsequent management of collected data can be a cumbersome activity that requires significant time and effort, from all involved parties. Tracking and managing student attendance during lectures and exams is an especially important instance of that task. Automated attendance management systems devised in the literature differ in applied technological solutions and features. However, many of these lack some features that would make them particularly useful for the task of student attendance management. This paper proposes the design of an automatic system tailored for the tracking and management of student attendance data. The design incorporates three main components that interact to deliver a seamless user experience. The key element is the utilisation of simple beacons, which make the system cost-effective and easy to use. Also, the design is open for future integration of sophisticated data analysis.
- 9:37 Modelling and Characterization of FBG Sensors
- In this paper, modelling and characterization of strain and temperature sensors based on fiber Bragg grating is presented. Simulation of FBG sensors behavior demonstrates very good measurement linearity for both strain and temperature. Simulation results were experimentally verified using two common methods for FBG interrogation. Measurement results obtained by both methods show a good match and good linearity. Strain sensitivity of 0.0002 nm/µε and temperature sensitivity of approximately 0.014 nm/˚C were obtained.
- 9:56 Review of automotive audio technologies: immersive audio case study
- Advances in autonomous cars are driving researches in different areas, including infotainment, electric cars acoustics and multimedia user experience. Lines between automotive and consumer multimedia technologies are not that strict anymore, and that opens lot of questions regarding platforms, interfaces, safety, technologies and technologies' updates which are much faster in consumer world. In this paper we review current automotive audio technologies and researches, and evaluate object based audio application on automotive platform. Objects based audio technologies are state of the art in home audio and theatres and next step in automotive multimedia.
Thursday, May 30 9:00 - 11:00
Workshop1: "Autonomous driving 101: the full story"
Autonomous driving 101: the full storyRoom: B1
Chairs: Nives Kaprocki, Goran Miljkovic
- 9:00 Autonomous driving 101: the full story
As a result of recent advances in many areas, such as large open source projects, technology has become far more accessible than ever before. Autonomous driving is no longer reserved only for the big players but can be applied by everyone to a certain extent. As a student team with years of experience in motorsports, the TUfast Eco team, therefore, decided to develop its very own autonomous vehicle. Based on the experiences with the new generation of this research vehicle, we will summarize important key points that are necessary for the realization of such a project. This includes an extensive overview of the hardware and software architecture required for a fully functional autonomous vehicle. The whole story of the development is being told, with all of the problems and their possible solutions. The aim of this talk is to give the audience an understanding of the big picture, from the processing of the sensor data to the control of the actuators. Thereby we look at the individual aspects and explain important concepts such as perception and path planning in an accessible manner. Furthermore, we demonstrate the necessity of simulation and the synthetic generation of training data.
Thursday, May 30 9:00 - 15:45
Workshop2: "Fundamentals of Deep Learning for Computer Vision"
Fundamentals of Deep Learning for Computer VisionRoom: C1
Chair: Milan Z. Bjelica
- 9:00 Fundamentals of Deep Learning for Computer Vision
Fundamentals of Deep Learning for Computer Vision.
Thursday, May 30 10:15 - 11:00
Keynote5: "Employee wellbeing in the digital age"
Employee wellbeing in the digital ageRoom: A1
Chairs: Branislav M. Todorović, Aleksa Zejak
- 10:15 Employee wellbeing in the digital age
- The accelerated development of the IT sector has been reconfiguring the labour market by creating new jobs and by changing work environments. It has linked people from all continents, allowing them to communicate easily within a virtual world. Even though IT has brought obvious benefits, it has also brought some serious challenges relate to employee wellbeing. In the last few years, we have faced some serious problems that emerged as a consequence of mismanaged technological interventions. They have led to blurring the line between work and private life. Employee availability has made never-ending workdays a fact of life. Some new work stressors, such as technostress and workplace cyberbullying, have emerged. Solving these problems is a momentous challenge for experts from various scientific fields. The scientific endeavours have to point to the ways in which technology could be brought back to the role of improving work performance in a way that would enhance human wellbeing and make employees enthusiastic about their future in the digital age.
Thursday, May 30 11:00 - 11:30
Galvanize your network: Coffee break
Rooms: A1, B1, Caffeteria
Chair: Jelena Gajic
Thursday, May 30 11:30 - 12:00
Keynote6: "How My Own IoT and Spyware Projects Drive My Family Mad"
How My Own IoT and Spyware Projects Drive My Family MadRoom: A1
Chairs: Lucio Ciabattoni, Ioan Tache
- 11:30 How My Own IoT and Spyware Projects Drive My Family Mad
- In 2014 I started developing my own IoT and Home Automation project. At the same time, I have developed a software that restricts time on my kids computers and mobile phones. Now I have a perfect line of spying projects, but I am the only one who loves them.
Thursday, May 30 11:30 - 12:45
Student session: Galvanizing Automotive solutions 2
Chairs: Tomislav Maruna, Vesna Petrović
- 11:30 Automotive Vision Grabber: FPGA design, cameras and data transfer over PCIe
- In this paper, a system which enables testing and faster development in automotive industry is proposed. The system enables testing in a very simple way, allowing cost savings during testing and development, as well as raising security in a way that first tests are not performed on real roads, but are done in the laboratory. Automotive Machine Vision (AMV) Grabber is a device whose purpose is capturing multiple video streams and saving them to hard disk drive. Captured streams could then be used for testing various Advanced Driver Assistance System (ADAS) algorithms used in automotive. This paper shows implementation of Field Programmable Gate Array (FPGA) design, camera control system and Peripheral Component Interconnect Express (PCIe) interface. Furthermore, different tests were performed, while the obtained results are promising.
- 11:48 Time to collision estimation for vehicles coming from behind using in-vehicle camera
- In recent years a number of Advanced Driver-Assistance Systems (ADAS) implemented in vehicles has been rising significantly. In this paper an algorithm for ADAS that estimates the time to potential collision (TTC) with the vehicle coming from behind, by processing only the video frames captured by the camera located in the outside rear-view mirror, is proposed. After preprocessing of the input video frame, vehicle detection is performed, followed by vehicle tracking. When the vehicle is successfully detected and tracked, it is possible to estimate the distance between oncoming vehicle and camera, as well as the oncoming vehicle speed. Using the calculated distance and the speed of the oncoming vehicle, the proposed algorithm outputs the final information about the TTC, i.e. time remaining for safe change of the driving lane. The algorithm has been tested on different real video sequences and it achieves high performance in terms of estimated TTC accuracy, while processing 25 Full HD frames per second.
- 12:07 Classification of Objects Detected by the Camera based on Convolutional Neural Network
- Nowadays, we are trying to achieve as much vehicle autonomy as possible by developing Advanced Driver-Assistance Systems (ADAS). For such a system to make decisions, it should have insight into the environment of the vehicle, e.g. the objects surrounding the vehicle. During forward driving, the information about the objects in front of the vehicle is usually obtained by a front view in-vehicle camera. This paper describes the image classification method of the objects in the front of the vehicle based on deep convolutional neural networks (CNN). Such CNN is supposed to be implemented in embedded system of an autonomous vehicle and the inference should satisfy real-time constraints. This means that the CNN should be structured to have fast inference by reducing the number of operations as much as possible, but still having satisfying accuracy. This can be achieved by reducing the number of parameters which also means that the resulting network has lower memory requirements. This paper describes the process of realizing such a network, from image dataset development up to the CNN structuring and training. The proposed CNN is compared to the state-of-the-art deep neural network in terms of classification accuracy, inference speed and memory requirements.
Thursday, May 30 12:00 - 13:00
Workshop3: "Large Scale Angular Architecture & Code Splitting Existing ones"
Large Scale Angular Architecture & Code Splitting Existing onesRoom: A1
Chairs: Vukota Pekovic, Ivan Velikic
- 12:00 Large Scale Angular Architecture & Code Splitting Existing ones
When designing multi-module, multi-app, distributed Angular apps, a common code-base and path aliasing is a must. We will go through best practices in these scenarios, although this is no silver bullet. As an example we will try to split an existing app into several, to define a shared code-base and publish it on a Gitlab/npm repository, so that each shared part of the code base can be referenced as a dependency.
Thursday, May 30 13:00 - 14:15
Rooms: A1, B1, Caffeteria
Chair: Jelena Gajic
Thursday, May 30 14:15 - 15:00
Keynote 7: "Cloud-based validation of automated driving"
Cloud-based validation of automated drivingRoom: A1
Chairs: Soumya Kanti Datta, Velibor Mihic
- 14:15 Cloud-based validation of automated driving
- Cloud-based validation of automated driving.
Thursday, May 30 14:15 - 15:30
Galvanizing AI, Big Data, Automotive, Branding
Chairs: Tihomir Andjelic, Milena Milosevic
- 14:15 Fuzzy analysis of surface roughness in electrical discharge machining
- The paper proposes an artificial intelligence tool, respectively fuzzy logic technique for analyzing surface roughness of electrical discharge machining. The analysis of surface roughness based on the fuzzy logic method, which is an approach for analysis of the electrical discharge machining was researched using the Taguchi L9 experimental plan. The objective of research is to examine the influence of the individual important machining parameters, like discharge current and pulse duration on surface roughness. The results indicate that the fuzzy logic technique can be effectively used for the analysis of surface roughness in electrical discharge machining.
- 14:33 Identification of Essential Patents in AI based Automotive Patent Portfolio
- With continuous development of automotive industry, a huge need for the analysis of patents that covers the key technologies applied in the automotive industry is expected. According to this fact, the identification of essential patents of the automotive industry patent portfolio is of great importance. This paper deals mainly with a methodology for identifying essential patents in a way those emphases the analysis of technical aspects of a patent, and not so much the market side. Based on technical attributes obtained from patent documents, patents are grouped and processed, and at the end, a small percentage of patents with high score is selected and determined as essential.
- 14:52 Predicting Positions and Velocities of Surrounding Vehicles using Deep Neural Networks
- Prediction of surrounding vehicles motion is basic feature of the most advanced driver assistance systems (ADAS). In this paper, we present prediction of position and velocity of surrounding vehicles using deep neural networks (DNNs). Three different DNN architectures are designed and explored: feed-forward, recurrent, and hybrid. Training and validation data are generated using IPG Carmaker simulation environment. The reliability and accuracy of prediction models under simulated highway conditions and variable number of input-output time steps have been examined. Hybrid DNN showed better performance compared to feed-forward and recurrent neural networks.
- 15:11 An Automated Procedure to Evaluate Usability of eHealth Platforms In The Wild
- Nowadays, one of the keys for the success of ICT technologies, in particular in the eHealth realm, is "Usability" and self-assessment questionnaires are the most diffuse procedure for its evaluation. However, this procedure presents different limitations (e.g., subjectivity, needs of expert supervisors, etc.). In this paper a methodology, based on emotion recognition, web analytics and gaze detection, is developed in order to asses usability of eHealth web platforms objectively, automatically, remotely and in the wild.
Thursday, May 30 15:00 - 15:45
Keynote8: "AMICA: Activity Monitoring and Coaching Application for IoT Enabled Healthcare"
AMICA: Activity Monitoring and Coaching Application for IoT Enabled HealthcareRoom: A1
Chairs: Gabriele Foresi, Boris Radin
- 15:00 AMICA: Activity Monitoring and Coaching Application for IoT Enabled Healthcare
- AMICA: Activity Monitoring and Coaching Application for IoT Enabled Healthcare.
Thursday, May 30 15:45 - 16:15
Galvanize your network: Coffee break
Rooms: A1, B1, Caffeteria
Chair: Jelena Gajic
Thursday, May 30 16:15 - 17:00
Awards & Farewell Until our Roads Meet Again
Chair: Milan Vidakovic