ADAS audio automotive availability estimation business cloud computer vision consumer electronics context-aware deep learning digital TV functional safety gaming HMI human detection ieeexplore image processing infotainment internet of things keynote lighting control market research networks RSSI smart homes software framework user experience user interface user monitoring zigbee
21 entries (M: 33.5)
2024 |
Bjelica, Milan Z Designing a safe autonomous vehicle computer - where we are, where we should be and a hands-on example ConferenceM32 2024 IEEE International Conference on Consumer Electronics (ICCE), IEEE, 2024, ISBN: 979-8-3503-2413-6. Abstract | Links | BibTeX | Tags: ADAS, automotive, functional safety @conference{ICCE2024, title = {Designing a safe autonomous vehicle computer - where we are, where we should be and a hands-on example}, author = {Milan Z. Bjelica}, url = {https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10444258}, doi = {10.1109/ICCE59016.2024.10444258}, isbn = {979-8-3503-2413-6}, year = {2024}, date = {2024-01-05}, booktitle = {2024 IEEE International Conference on Consumer Electronics (ICCE)}, publisher = {IEEE}, abstract = {In the last two years, we have witnessed the increased push to legislators to approve the robotaxi vehicles in several cities in the USA for commercial use, with two permits Issued, and one revoked. The endeavor to proliferate self-driving vehicles proves to be very hard, since the technology stack required, dominated by software, presents extreme challenges In safety critical design and certification according to practices laid out in ISO 26262 and ISO 21448 SOTIF. In this tutorial we would contrast the due diligence in automotive functional safety with the real-world design challenges and what has been deployed on roads. We would discuss legal debates around the technology shortcomings following the first lawsuits regarding the casualties caused by the car autopilots. Finally, we would work out in a hands-on example a process of designing a safe autonomous vehicle computer for a traffic jam pilot function, witnessing all the challenges first hand.}, howpublished = {M32}, keywords = {ADAS, automotive, functional safety}, pubstate = {published}, tppubtype = {conference} } In the last two years, we have witnessed the increased push to legislators to approve the robotaxi vehicles in several cities in the USA for commercial use, with two permits Issued, and one revoked. The endeavor to proliferate self-driving vehicles proves to be very hard, since the technology stack required, dominated by software, presents extreme challenges In safety critical design and certification according to practices laid out in ISO 26262 and ISO 21448 SOTIF. In this tutorial we would contrast the due diligence in automotive functional safety with the real-world design challenges and what has been deployed on roads. We would discuss legal debates around the technology shortcomings following the first lawsuits regarding the casualties caused by the car autopilots. Finally, we would work out in a hands-on example a process of designing a safe autonomous vehicle computer for a traffic jam pilot function, witnessing all the challenges first hand. |
2022 |
Simić, Đorđe; Stefanović, Stefan; Dankulov, Marija Mitrović; Stepanenko, Dimitrije; Bjelica, Milan Z Autonomous mobility: appropriate tools and verification practices ConferenceM32 2022 IEEE Zooming Innovation in Consumer Technologies Conference (ZINC), 2022, ISBN: 978-1-6654-8374-2. Abstract | Links | BibTeX | Tags: ADAS, automotive @conference{BjelicaZinc2022, title = {Autonomous mobility: appropriate tools and verification practices}, author = {Đorđe Simić and Stefan Stefanović and Marija Mitrović Dankulov and Dimitrije Stepanenko and Milan Z. Bjelica}, doi = {10.1109/ZINC55034.2022.9840668}, isbn = {978-1-6654-8374-2}, year = {2022}, date = {2022-05-25}, booktitle = {2022 IEEE Zooming Innovation in Consumer Technologies Conference (ZINC)}, abstract = {A panel of distinguished speakers from both industry and academia, discuss autonomous mobility “from within”. What is required to develop autonomous mobility solutions? Which algorithms are needed? Can we rely upon those solutions in realistic environments? Can we simulate the real world so that we can perform adequate laboratory pre-verification and validation? Institute of Physics Belgrade, SYRMIA LLC, and University of California San Diego discuss together with session presenters to answer the questions above!}, howpublished = {M32}, keywords = {ADAS, automotive}, pubstate = {published}, tppubtype = {conference} } A panel of distinguished speakers from both industry and academia, discuss autonomous mobility “from within”. What is required to develop autonomous mobility solutions? Which algorithms are needed? Can we rely upon those solutions in realistic environments? Can we simulate the real world so that we can perform adequate laboratory pre-verification and validation? Institute of Physics Belgrade, SYRMIA LLC, and University of California San Diego discuss together with session presenters to answer the questions above! |
2020 |
Manić, Milan; Bjelica, Milan Z; Pešić, Jasmina Proposal for in-car driver alerting system of obstacles and surrounding vehicles ConferenceM33 2020 28th Telecommunications Forum (TELFOR), IEEE, 2020, ISBN: 978-1-6654-0499-0. Abstract | Links | BibTeX | Tags: ADAS, automotive, HMI @conference{2020tfa, title = {Proposal for in-car driver alerting system of obstacles and surrounding vehicles}, author = {Milan Manić and Milan Z. Bjelica and Jasmina Pešić}, doi = {10.1109/TELFOR51502.2020.9306515}, isbn = {978-1-6654-0499-0}, year = {2020}, date = {2020-11-24}, booktitle = {2020 28th Telecommunications Forum (TELFOR)}, publisher = {IEEE}, abstract = {In the automotive industry, safety is a very big issue and there are various driver assistance systems and driver warning systems. Many of these solutions are based on a software system to alert you with audio or visual warning messages. The problem with these solutions is when the driver cannot see or hear these warnings and therefore cannot react promptly. This paper presents a solution on how to warn the driver of the presence of obstacles or other cars in the immediate vicinity by using vibrations as a medium for transmitting messages between the car and the driver. Objects in the immediate vicinity are detected by various sensors and cameras, and then this information is used to make decisions about the nature of the alerts to be activated. In this paper, the hardware and software for the realization of this solution are presented. The system consists of a vibrating seat, a sensor (camera), a controller and a software solution. The principle of operation is as follows: when there is an obstacle on the right side of the vehicle, such as a passing car or some other obstacle, the seat will vibrate on the right side to inform the driver about the existence of an obstacle. The intensity of the vibrations corresponds to the distance from the obstacle, the stronger vibrations correspond to the object that is closer to the car. Thus, vibrations will alert the driver to the presence and distance of an obstacle or a passing car.}, howpublished = {M33}, keywords = {ADAS, automotive, HMI}, pubstate = {published}, tppubtype = {conference} } In the automotive industry, safety is a very big issue and there are various driver assistance systems and driver warning systems. Many of these solutions are based on a software system to alert you with audio or visual warning messages. The problem with these solutions is when the driver cannot see or hear these warnings and therefore cannot react promptly. This paper presents a solution on how to warn the driver of the presence of obstacles or other cars in the immediate vicinity by using vibrations as a medium for transmitting messages between the car and the driver. Objects in the immediate vicinity are detected by various sensors and cameras, and then this information is used to make decisions about the nature of the alerts to be activated. In this paper, the hardware and software for the realization of this solution are presented. The system consists of a vibrating seat, a sensor (camera), a controller and a software solution. The principle of operation is as follows: when there is an obstacle on the right side of the vehicle, such as a passing car or some other obstacle, the seat will vibrate on the right side to inform the driver about the existence of an obstacle. The intensity of the vibrations corresponds to the distance from the obstacle, the stronger vibrations correspond to the object that is closer to the car. Thus, vibrations will alert the driver to the presence and distance of an obstacle or a passing car. |
Manić, Milan; Bjelica, Milan Z Proposal for visual warnings system in the automotive digital cockpit using graphics sharing ConferenceM33 2020 28th Telecommunications Forum (TELFOR), IEEE, 2020. Abstract | Links | BibTeX | Tags: ADAS, automotive, HMI, ieeexplore, infotainment @conference{2020tfb, title = {Proposal for visual warnings system in the automotive digital cockpit using graphics sharing}, author = {Milan Manić and Milan Z. Bjelica}, doi = {10.1109/TELFOR51502.2020.9306638}, year = {2020}, date = {2020-11-24}, booktitle = {2020 28th Telecommunications Forum (TELFOR)}, publisher = {IEEE}, abstract = {In this paper is presented the concept of using shared graphics to display visual warnings in the automotive digital cockpit. Given that cars today have over 100 ECUs (ECU - control unit), communication and management of all these ECUs become a challenging task. This paper focuses on the control units used to display the content and process the data of each screen in the car. The number of control units for data processing and display of content can be reduced by using SoC (SoC - System on a chip) with a hypervisor. A hypervisor is a concept that allows us to run two operating systems on one SoC in real-time. The proposed system consists of one SoC with two operating systems running on a hypervisor and having the ability to display content on three different screens. The proposed solution covers the simultaneous display of content from both operating systems on one screen as well as the display of visual warnings on all screens in the digital cockpit, regardless of the operating system that hosts that screen. Alerts are displayed from a higher security operating system regardless of the content currently displayed from a lower security operating system. This approach has led to the certification of operating systems in terms of security.}, howpublished = {M33}, keywords = {ADAS, automotive, HMI, ieeexplore, infotainment}, pubstate = {published}, tppubtype = {conference} } In this paper is presented the concept of using shared graphics to display visual warnings in the automotive digital cockpit. Given that cars today have over 100 ECUs (ECU - control unit), communication and management of all these ECUs become a challenging task. This paper focuses on the control units used to display the content and process the data of each screen in the car. The number of control units for data processing and display of content can be reduced by using SoC (SoC - System on a chip) with a hypervisor. A hypervisor is a concept that allows us to run two operating systems on one SoC in real-time. The proposed system consists of one SoC with two operating systems running on a hypervisor and having the ability to display content on three different screens. The proposed solution covers the simultaneous display of content from both operating systems on one screen as well as the display of visual warnings on all screens in the digital cockpit, regardless of the operating system that hosts that screen. Alerts are displayed from a higher security operating system regardless of the content currently displayed from a lower security operating system. This approach has led to the certification of operating systems in terms of security. |
Jelić, Borna; Grbić, Ratko; Vranješ, Mario; Bjelica, Milan Z UrTra2D – Urban Traffic 2D Object Detection Dataset ConferenceM33 2020 IEEE 10th International Conference on Consumer Electronics (ICCE-Berlin), 2020, ISBN: 978-1-7281-5885-3. Abstract | Links | BibTeX | Tags: ADAS, automotive, deep learning, ieeexplore @conference{2020berlin1, title = {UrTra2D – Urban Traffic 2D Object Detection Dataset}, author = {Borna Jelić and Ratko Grbić and Mario Vranješ and Milan Z. Bjelica}, doi = {10.1109/ICCE-Berlin50680.2020.9352154}, isbn = {978-1-7281-5885-3}, year = {2020}, date = {2020-11-09}, booktitle = {2020 IEEE 10th International Conference on Consumer Electronics (ICCE-Berlin)}, abstract = {With progress being made in the field of artificial intelligence and especially machine learning, tech and vehicle companies acquired a powerful tool and made a large step towards realisation of a fully autonomous vehicle. Along with the exploding development of more and more powerful hardware, deep learning has become one of the most dominant fields of research in the automotive domain, succeeding the classical computer vision methods. However, to be able to apply deep learning methods to solve a problem, large and appropriate datasets are required in developing a solution, as there is never enough data for deep learning. In this paper, Urban Traffic 2D Object Detection (UrTra2D) dataset is presented, which is intended for training 2D detectors of specific objects common for urban traffic scenes. The data was recorded with an affordable camera mounted inside the vehicle. The dataset contains video sequences and labelled frames of the traffic in the city of Osijek in different weather conditions during both day and night. There are 5 770 labelled frames, totalling in 22 764 labelled objects throughout 11 categories. The UrTra2D dataset is freely available to the research community upon request.}, howpublished = {M33}, keywords = {ADAS, automotive, deep learning, ieeexplore}, pubstate = {published}, tppubtype = {conference} } With progress being made in the field of artificial intelligence and especially machine learning, tech and vehicle companies acquired a powerful tool and made a large step towards realisation of a fully autonomous vehicle. Along with the exploding development of more and more powerful hardware, deep learning has become one of the most dominant fields of research in the automotive domain, succeeding the classical computer vision methods. However, to be able to apply deep learning methods to solve a problem, large and appropriate datasets are required in developing a solution, as there is never enough data for deep learning. In this paper, Urban Traffic 2D Object Detection (UrTra2D) dataset is presented, which is intended for training 2D detectors of specific objects common for urban traffic scenes. The data was recorded with an affordable camera mounted inside the vehicle. The dataset contains video sequences and labelled frames of the traffic in the city of Osijek in different weather conditions during both day and night. There are 5 770 labelled frames, totalling in 22 764 labelled objects throughout 11 categories. The UrTra2D dataset is freely available to the research community upon request. |
2019 |
Bjelica, Milan Z; Lukač, Željko Central vehicle computer design: Software taking over Journal ArticleM22 IEEE Consumer Electronics Magazine, 8 (6), 2019, ISSN: 2162-2248. Abstract | Links | BibTeX | Tags: ADAS, automotive, ieeexplore, infotainment, market research, software framework @article{cemag_2019, title = {Central vehicle computer design: Software taking over}, author = {Milan Z. Bjelica and Željko Lukač}, doi = {10.1109/MCE.2019.2929813}, issn = {2162-2248}, year = {2019}, date = {2019-11-01}, journal = {IEEE Consumer Electronics Magazine}, volume = {8}, number = {6}, abstract = {To prevent each vehicle manufacturer from “reinventing the wheel” and spending vast amounts of time and engineering, a paradigm shift is needed. Transferring the vehicle design to standardized hardware would enable software to take the key role and allow the horizontal approach in design, where each feature may be added as a module. This sets the stage for a central vehicle computer-a brain for next generation vehicles which is everything but easy to design. This article discusses one such design and identifies the required building blocks for this rising market.}, howpublished = {M22}, keywords = {ADAS, automotive, ieeexplore, infotainment, market research, software framework}, pubstate = {published}, tppubtype = {article} } To prevent each vehicle manufacturer from “reinventing the wheel” and spending vast amounts of time and engineering, a paradigm shift is needed. Transferring the vehicle design to standardized hardware would enable software to take the key role and allow the horizontal approach in design, where each feature may be added as a module. This sets the stage for a central vehicle computer-a brain for next generation vehicles which is everything but easy to design. This article discusses one such design and identifies the required building blocks for this rising market. |
Bjelica, Milan Z; Marinković, Vladimir; Đukić, Miodrag; Lukač, Željko The system and method for real-time decision-making during autonomous driving PatentPendingM87 P-2019/0838, 2019, (Pending). BibTeX | Tags: ADAS, automotive @patent{2019p1, title = {The system and method for real-time decision-making during autonomous driving}, author = {Milan Z. Bjelica and Vladimir Marinković and Miodrag Đukić and Željko Lukač}, year = {2019}, date = {2019-11-01}, number = {P-2019/0838}, howpublished = {M87}, note = {Pending}, keywords = {ADAS, automotive}, pubstate = {published}, tppubtype = {patent} } |
Bjelica, Milan Z; Marinković, Vladimir; Đukić, Miodrag; Kaštelan, Ivan A system of software components for isolated execution of an artificial intelligence algorithm for vehicle PatentPendingM87 P-2019/1098, 2019, (Pending). BibTeX | Tags: ADAS, automotive, deep learning, software framework @patent{2019p2, title = {A system of software components for isolated execution of an artificial intelligence algorithm for vehicle}, author = {Milan Z. Bjelica and Vladimir Marinković and Miodrag Đukić and Ivan Kaštelan}, year = {2019}, date = {2019-10-01}, number = {P-2019/1098}, howpublished = {M87}, note = {Pending}, keywords = {ADAS, automotive, deep learning, software framework}, pubstate = {published}, tppubtype = {patent} } |
Milošević, Marko; Isić, Vesna; Bjelica, Milan Z; Anđelić, Tihomir Efficient Implementation of Camera Mirror System Algorithm on Heterogeneous Chip Architectures ConferenceM33 Consumer Electronics - Berlin (ICCE-Berlin), 2019 IEEE 9th International Conference on, IEEE Consumer Electronics Society, 2019, ISBN: 978-1-7281-2745-3. Abstract | Links | BibTeX | Tags: ADAS, automotive, ieeexplore @conference{icceberlin2019_1, title = {Efficient Implementation of Camera Mirror System Algorithm on Heterogeneous Chip Architectures}, author = {Marko Milošević and Vesna Isić and Milan Z. Bjelica and Tihomir Anđelić}, doi = {10.1109/ICCE-Berlin47944.2019.8966193}, isbn = {978-1-7281-2745-3}, year = {2019}, date = {2019-09-08}, booktitle = {Consumer Electronics - Berlin (ICCE-Berlin), 2019 IEEE 9th International Conference on}, publisher = {IEEE Consumer Electronics Society}, abstract = {Advanced Driving Assistance Systems (ADAS) are vehicle-based intelligent safety systems which help the drivers to improve safety driving. One important building block of ADAS is a Camera Mirror System (CMS). CMS provides means to enhance rearview mirrors by digital information, such as detected obstacles, collision avoidance alerts and more. Efficient implementation of CMS requires the utilization of dedicated processing in hardware, to ensure minimal latency and coexistence with other algorithms in heterogeneous hardware environments. In this paper we analyze CMS algorithm data pipeline and provide insights on how each phase can be efficiently accelerated using dedicated hardware blocks on present-day high-performance microcontrollers. We give early evaluation of the presented approach on top of two architectures: TI TDA2x and NVIDIA Xavier.}, howpublished = {M33}, keywords = {ADAS, automotive, ieeexplore}, pubstate = {published}, tppubtype = {conference} } Advanced Driving Assistance Systems (ADAS) are vehicle-based intelligent safety systems which help the drivers to improve safety driving. One important building block of ADAS is a Camera Mirror System (CMS). CMS provides means to enhance rearview mirrors by digital information, such as detected obstacles, collision avoidance alerts and more. Efficient implementation of CMS requires the utilization of dedicated processing in hardware, to ensure minimal latency and coexistence with other algorithms in heterogeneous hardware environments. In this paper we analyze CMS algorithm data pipeline and provide insights on how each phase can be efficiently accelerated using dedicated hardware blocks on present-day high-performance microcontrollers. We give early evaluation of the presented approach on top of two architectures: TI TDA2x and NVIDIA Xavier. |
Baba, Filip; Kenjić, Dušan; Bjelica, Milan Z; Kaštelan, Ivan Optimizing Deep Learning Based Semantic Video Segmentation on Embedded GPUs ConferenceM33 Consumer Electronics - Berlin (ICCE-Berlin), 2019 IEEE 9th International Conference on, IEEE, 2019, ISBN: 978-1-7281-2745-3. Abstract | Links | BibTeX | Tags: ADAS, automotive, deep learning, ieeexplore @conference{icceberlin2019_2, title = {Optimizing Deep Learning Based Semantic Video Segmentation on Embedded GPUs}, author = {Filip Baba and Dušan Kenjić and Milan Z. Bjelica and Ivan Kaštelan}, doi = {10.1109/ICCE-Berlin47944.2019.8966156}, isbn = {978-1-7281-2745-3}, year = {2019}, date = {2019-09-08}, booktitle = {Consumer Electronics - Berlin (ICCE-Berlin), 2019 IEEE 9th International Conference on}, publisher = {IEEE}, abstract = {Decision making in many industries today is being improved drastically thanks to artificial intelligence and deep learning. New algorithms address challenges such as genome mapping, medical diagnostics, self-driving cars, autonomous robots and more. Deep learning in embedded systems requires high optimization due to the high computational demand, given that power, heat dissipation, size and price constraints are numerous. In this paper we analyze several acceleration methods which include utilization of GPUs for most complex variants of deep learning, such as semantic video segmentation operating in real time. Specifically, we propose mapping of acceleration routines commonly present within deep learning SDKs to different network layers in semantic segmentation. Finally, we evaluate one implementation utilizing the enumerated techniques for semantic segmentation of front camera in autonomous driving front view.}, howpublished = {M33}, keywords = {ADAS, automotive, deep learning, ieeexplore}, pubstate = {published}, tppubtype = {conference} } Decision making in many industries today is being improved drastically thanks to artificial intelligence and deep learning. New algorithms address challenges such as genome mapping, medical diagnostics, self-driving cars, autonomous robots and more. Deep learning in embedded systems requires high optimization due to the high computational demand, given that power, heat dissipation, size and price constraints are numerous. In this paper we analyze several acceleration methods which include utilization of GPUs for most complex variants of deep learning, such as semantic video segmentation operating in real time. Specifically, we propose mapping of acceleration routines commonly present within deep learning SDKs to different network layers in semantic segmentation. Finally, we evaluate one implementation utilizing the enumerated techniques for semantic segmentation of front camera in autonomous driving front view. |
Gamf, Branislav; Usorac, Srđan; Bjelica, Milan Z; Lukač, Željko Video delivery subsystem for multi-SoC automotive machine vision platforms ConferenceM33 Consumer Electronics - Berlin (ICCE-Berlin), 2019 IEEE 9th International Conference on, IEEE, 2019, ISBN: 978-1-7281-2745-3. Abstract | Links | BibTeX | Tags: ADAS, automotive, ieeexplore @conference{Gamf2019, title = {Video delivery subsystem for multi-SoC automotive machine vision platforms}, author = {Branislav Gamf and Srđan Usorac and Milan Z. Bjelica and Željko Lukač}, doi = {10.1109/ICCE-Berlin47944.2019.8966206}, isbn = {978-1-7281-2745-3}, year = {2019}, date = {2019-09-08}, booktitle = {Consumer Electronics - Berlin (ICCE-Berlin), 2019 IEEE 9th International Conference on}, publisher = {IEEE}, abstract = {Automotive microcontrollers in a form of System-on-Chip (SoC) are getting more powerful due to the abundance of algorithms catering to the Advanced Driver-Assistance System (ADAS) applications. Apart from traditional LIDAR and RADAR-based approaches, many new algorithms heavily depend on cameras and video. Delivering video within a heterogeneous hardware environment such as within a single SoC or across SoCs is a daunting task. In this paper, we give a proposal of a video delivery subsystem to be applied to next generation hardware and software architectures for autonomous vehicles. Proposed subsystem considers all major video exchange routes, including DMA, PCI-E and Ethernet, with appropriate software interfaces which can be integrated to future horizontal automotive middleware. Within an early evaluation we demonstrate the utilization of the subsystem in an automotive hardware encompassing three TI TDA2X and two NVIDIA Xavier SoCs.}, howpublished = {M33}, keywords = {ADAS, automotive, ieeexplore}, pubstate = {published}, tppubtype = {conference} } Automotive microcontrollers in a form of System-on-Chip (SoC) are getting more powerful due to the abundance of algorithms catering to the Advanced Driver-Assistance System (ADAS) applications. Apart from traditional LIDAR and RADAR-based approaches, many new algorithms heavily depend on cameras and video. Delivering video within a heterogeneous hardware environment such as within a single SoC or across SoCs is a daunting task. In this paper, we give a proposal of a video delivery subsystem to be applied to next generation hardware and software architectures for autonomous vehicles. Proposed subsystem considers all major video exchange routes, including DMA, PCI-E and Ethernet, with appropriate software interfaces which can be integrated to future horizontal automotive middleware. Within an early evaluation we demonstrate the utilization of the subsystem in an automotive hardware encompassing three TI TDA2X and two NVIDIA Xavier SoCs. |
Bjelica, Milan Z Deep Learning vs. Safety - Practical Approach and Platform Design Perspective ConferenceKeynoteM32 Proceedings of 2019 International Conference on Systems, Signals and Image Processing (IWSSIP), EURASIP, Osijek, Croatia, 2019, (Keynote). Abstract | Links | BibTeX | Tags: ADAS, automotive, deep learning, keynote, market research @conference{2019_iwssip, title = {Deep Learning vs. Safety - Practical Approach and Platform Design Perspective}, author = {Milan Z. Bjelica}, url = {http://www.milanbjelica.info/wp-content/uploads/2019/06/program-web.pdf https://www.youtube.com/watch?v=tJrDJsBiDqg}, year = {2019}, date = {2019-06-06}, booktitle = {Proceedings of 2019 International Conference on Systems, Signals and Image Processing (IWSSIP)}, publisher = {EURASIP}, address = {Osijek, Croatia}, abstract = {Deep Learning is a promising field, allowing an increase in artificial intelligence applications across many fields, ranging from data science, medical, weather, and aerospace to automotive. Applications of computer vision-based deep learning are vastly assisted by modern System-on-Chip architectures, which provide the required parallelism, heterogeneity and interfacing. However, the application of deep learning to safety-critical contexts where human lives might be at stake, such as in self-driving cars, still has many pitfalls. Ongoing academic research tackles transparent AI, in which the correctness of AI is attempted to be reached by design; however, the outcome of this research is still far-fetched. In this talk, we will discuss a practical approach when integrating deep learning vision-based solutions into a safety-critical context, which can be achieved today. We outline an approach which introduces a software/hardware platform design which fosters diversity, with the goal of minimizing risk of critical failures which are induced by AI in decision making.}, howpublished = {M32}, note = {Keynote}, keywords = {ADAS, automotive, deep learning, keynote, market research}, pubstate = {published}, tppubtype = {conference} } Deep Learning is a promising field, allowing an increase in artificial intelligence applications across many fields, ranging from data science, medical, weather, and aerospace to automotive. Applications of computer vision-based deep learning are vastly assisted by modern System-on-Chip architectures, which provide the required parallelism, heterogeneity and interfacing. However, the application of deep learning to safety-critical contexts where human lives might be at stake, such as in self-driving cars, still has many pitfalls. Ongoing academic research tackles transparent AI, in which the correctness of AI is attempted to be reached by design; however, the outcome of this research is still far-fetched. In this talk, we will discuss a practical approach when integrating deep learning vision-based solutions into a safety-critical context, which can be achieved today. We outline an approach which introduces a software/hardware platform design which fosters diversity, with the goal of minimizing risk of critical failures which are induced by AI in decision making. |
2018 |
Milošević, Milena; Bjelica, Milan Z; Maruna, Tomislav; Teslić, Nikola Software Platform for Heterogeneous In-Vehicle Environments Journal ArticleM22 IEEE Transactions on Consumer Electronics, 64 (2), pp. 213-221, 2018, ISSN: 0098-3063. Abstract | Links | BibTeX | Download | Tags: ADAS, automotive, ieeexplore, software framework @article{Milosevic, title = {Software Platform for Heterogeneous In-Vehicle Environments}, author = {Milena Milošević and Milan Z. Bjelica and Tomislav Maruna and Nikola Teslić}, url = {http://www.milanbjelica.info/index.php/sdm_downloads/software-platform-for-heterogeneous-in-vehicle-environments/, Download}, doi = {10.1109/TCE.2018.2844737}, issn = {0098-3063}, year = {2018}, date = {2018-06-07}, journal = {IEEE Transactions on Consumer Electronics}, volume = {64}, number = {2}, pages = {213-221}, abstract = {Modern technologies lead to more sophisticated hardware, while software is becoming more complex. These trends are widely present in consumer electronics and do not bypass automotive electronics either. There is an evident recent growth in in-vehicle infotainment, telematics, advanced driver assistance systems (ADASs) and cluster development. The number of electronic control units (ECUs) in vehicle constantly grows. Since typical vehicle ECU is providing one function per vehicle, it becomes harder for manufacturers to manage these ECUs due to diverse nature of the system, hence a rising demand for ECU consolidation exists. With the availability of sophisticated hardware, powerful system-on-chips (SoCs) can be used for multiple functions inside a vehicle. The transition toward less ECUs is an ongoing process, in which software needs to be aligned first and then transferred to the same SoC. This paper presents the software platform for heterogeneous immersive in-vehicle environments, providing a step in software consolidation, by allowing same abstractions for diverse applications executing on various hardware platforms. It proposes a framework for the scalable development of ADAS from consumer level to different automotive safety levels, provides unified access toward algorithm building blocks, multi-sensor real-time environment and easy integration of algorithms, thus enabling shorter development time.}, howpublished = {M22}, keywords = {ADAS, automotive, ieeexplore, software framework}, pubstate = {published}, tppubtype = {article} } Modern technologies lead to more sophisticated hardware, while software is becoming more complex. These trends are widely present in consumer electronics and do not bypass automotive electronics either. There is an evident recent growth in in-vehicle infotainment, telematics, advanced driver assistance systems (ADASs) and cluster development. The number of electronic control units (ECUs) in vehicle constantly grows. Since typical vehicle ECU is providing one function per vehicle, it becomes harder for manufacturers to manage these ECUs due to diverse nature of the system, hence a rising demand for ECU consolidation exists. With the availability of sophisticated hardware, powerful system-on-chips (SoCs) can be used for multiple functions inside a vehicle. The transition toward less ECUs is an ongoing process, in which software needs to be aligned first and then transferred to the same SoC. This paper presents the software platform for heterogeneous immersive in-vehicle environments, providing a step in software consolidation, by allowing same abstractions for diverse applications executing on various hardware platforms. It proposes a framework for the scalable development of ADAS from consumer level to different automotive safety levels, provides unified access toward algorithm building blocks, multi-sensor real-time environment and easy integration of algorithms, thus enabling shorter development time. |
Bjelica, Milan Z; Lukač, Željko; Maruna, Tomislav; Mihić, Velibor System for processing graphic content of the digital video cockpit of the vehicles with separate operating controls PatentPendingM87 P-2018/0671, 2018, (Pending). BibTeX | Tags: ADAS, automotive, infotainment @patent{Bjelica2018Pat3, title = {System for processing graphic content of the digital video cockpit of the vehicles with separate operating controls}, author = {Milan Z. Bjelica and Željko Lukač and Tomislav Maruna and Velibor Mihić}, year = {2018}, date = {2018-05-01}, number = {P-2018/0671}, howpublished = {M87}, note = {Pending}, keywords = {ADAS, automotive, infotainment}, pubstate = {published}, tppubtype = {patent} } |
Bjelica, Milan Z; Papp, Ištvan; Lukač, Željko; Milošević, Milena The procedure for updating the software in the car using proxies PatentPendingM87 P-2018/0553, 2018, (Pending). BibTeX | Tags: ADAS, automotive @patent{BjelicaPat20182, title = {The procedure for updating the software in the car using proxies }, author = {Milan Z. Bjelica and Ištvan Papp and Željko Lukač and Milena Milošević}, year = {2018}, date = {2018-04-02}, number = {P-2018/0553}, howpublished = {M87}, note = {Pending}, keywords = {ADAS, automotive}, pubstate = {published}, tppubtype = {patent} } |
Bjelica, Milan Z; Velikić, Gordana; Marinković, Vladimir; Pjevalica, Nebojša The procedure for determining the cumulative risk level for an early warning system in vehicles PatentPendingM87 P-2018/0382, 2018, (Pending). BibTeX | Tags: ADAS, ASIL, automotive @patent{BjelicaPat2018, title = {The procedure for determining the cumulative risk level for an early warning system in vehicles}, author = {Milan Z. Bjelica and Gordana Velikić and Vladimir Marinković and Nebojša Pjevalica}, year = {2018}, date = {2018-03-01}, number = {P-2018/0382}, howpublished = {M87}, note = {Pending}, keywords = {ADAS, ASIL, automotive}, pubstate = {published}, tppubtype = {patent} } |
Bjelica, Milan Z; Teslić, Nikola; Milošević, Milena; Kovačević, Branimir The procedure for configuring the autonomous vehicle in-vehicle system PatentPendingM87 P-2018/0001, 2018, (Pending). BibTeX | Tags: ADAS, automotive @patent{Bjelica2017b, title = {The procedure for configuring the autonomous vehicle in-vehicle system }, author = {Milan Z. Bjelica and Nikola Teslić and Milena Milošević and Branimir Kovačević}, year = {2018}, date = {2018-01-01}, number = {P-2018/0001}, howpublished = {M87}, note = {Pending}, keywords = {ADAS, automotive}, pubstate = {published}, tppubtype = {patent} } |
2017 |
Bjelica, Milan Z; Marinković, Vladimir; Đukić, Miodrag; Kaštelan, Ivan The method for adaptive video transmission with cameras in a vehicle PatentM94 P-2017-1119, 2017. BibTeX | Tags: ADAS, automotive, image processing @patent{Bjelica2017b, title = {The method for adaptive video transmission with cameras in a vehicle}, author = {Milan Z. Bjelica and Vladimir Marinković and Miodrag Đukić and Ivan Kaštelan}, year = {2017}, date = {2017-12-01}, number = {P-2017-1119}, howpublished = {M94}, keywords = {ADAS, automotive, image processing}, pubstate = {published}, tppubtype = {patent} } |
Gojak, Veselin; Janjatović, Joakim; Vukota, Nataša; Milošević, Milena; Bjelica, Milan Z Informational bird's eye view system for parking assistance ConferenceM33 Consumer Electronics - Berlin (ICCE-Berlin), 2017 IEEE 7th International Conference on, IEEE, 2017, ISBN: 978-1-5090-4014-8. Abstract | Links | BibTeX | Tags: ADAS, automotive, ieeexplore, infotainment @conference{Gojak2017, title = {Informational bird's eye view system for parking assistance}, author = {Veselin Gojak and Joakim Janjatović and Nataša Vukota and Milena Milošević and Milan Z. Bjelica}, doi = {10.1109/ICCE-Berlin.2017.8210604}, isbn = {978-1-5090-4014-8}, year = {2017}, date = {2017-09-03}, booktitle = {Consumer Electronics - Berlin (ICCE-Berlin), 2017 IEEE 7th International Conference on}, publisher = {IEEE}, abstract = {Primary goals of Advanced Driver Assistance Systems (ADAS) are increased safety and situations with no stress for all traffic participants. Parking safety is improved by presenting real-time vehicle environment to the driver (surround view, bird's eye view). Bird's eye view is 3D camera and parking sensors system which assist driver with parking. It covers all angles and alerts the driver about obstacles, thus minimizing possibility of traffic accidents. This paper presents one portable informational bird's eye view system.}, howpublished = {M33}, keywords = {ADAS, automotive, ieeexplore, infotainment}, pubstate = {published}, tppubtype = {conference} } Primary goals of Advanced Driver Assistance Systems (ADAS) are increased safety and situations with no stress for all traffic participants. Parking safety is improved by presenting real-time vehicle environment to the driver (surround view, bird's eye view). Bird's eye view is 3D camera and parking sensors system which assist driver with parking. It covers all angles and alerts the driver about obstacles, thus minimizing possibility of traffic accidents. This paper presents one portable informational bird's eye view system. |
2016 |
Simić, Aleksandra; Kocić, Ognjen; Bjelica, Milan Z; Milošević, Milena Driver monitoring algorithm for Advanced Driver Assistance Systems ConferenceM33 Telecommunications Forum (TELFOR), 2016 24th, IEEE, Belgrade, Serbia, 2016, ISBN: 978-1-5090-4086-5. Abstract | Links | BibTeX | Download | Tags: ADAS, automotive, computer vision, driver monitoring, ieeexplore, image processing @conference{telfor1, title = {Driver monitoring algorithm for Advanced Driver Assistance Systems}, author = {Aleksandra Simić and Ognjen Kocić and Milan Z. Bjelica and Milena Milošević}, doi = {10.1109/TELFOR.2016.7818908}, isbn = {978-1-5090-4086-5}, year = {2016}, date = {2016-11-22}, booktitle = {Telecommunications Forum (TELFOR), 2016 24th}, publisher = {IEEE}, address = {Belgrade, Serbia}, abstract = {Fast expansion of Advanced Driver Assistance Systems (ADAS) market and applications has resulted in a high demand for various accompanying algorithms. In this paper we present an implementation of Driver monitoring algorithm. Main goal of the algorithm is to automatically asses if driver is tired and in that case, raise a proper alert. It is widely used as a standard component of rest recommendation systems. Our approach is based on combination of computer vision algorithms for face detection and eyes detection. Additionally, we have tested our implementation in controlled environment on a real ADAS platform board.}, howpublished = {M33}, keywords = {ADAS, automotive, computer vision, driver monitoring, ieeexplore, image processing}, pubstate = {published}, tppubtype = {conference} } Fast expansion of Advanced Driver Assistance Systems (ADAS) market and applications has resulted in a high demand for various accompanying algorithms. In this paper we present an implementation of Driver monitoring algorithm. Main goal of the algorithm is to automatically asses if driver is tired and in that case, raise a proper alert. It is widely used as a standard component of rest recommendation systems. Our approach is based on combination of computer vision algorithms for face detection and eyes detection. Additionally, we have tested our implementation in controlled environment on a real ADAS platform board. |
Kocić, Ognjen; Simić, Aleksandra; Bjelica, Milan Z; Maruna, Tomislav Optimization of driver monitoring ADAS algorithm for heterogeneous platform ConferenceM33 Telecommunications Forum (TELFOR), 2016 24th, IEEE, Belgrade, Serbia, 2016, ISBN: 978-1-5090-4086-5. Abstract | Links | BibTeX | Download | Tags: ADAS, automotive, driver monitoring, ieeexplore @conference{telfor2, title = {Optimization of driver monitoring ADAS algorithm for heterogeneous platform}, author = {Ognjen Kocić and Aleksandra Simić and Milan Z. Bjelica and Tomislav Maruna}, doi = {10.1109/TELFOR.2016.7818910}, isbn = {978-1-5090-4086-5}, year = {2016}, date = {2016-11-22}, booktitle = {Telecommunications Forum (TELFOR), 2016 24th}, publisher = {IEEE}, address = {Belgrade, Serbia}, abstract = {Rapid expansion of Advanced Driver Assistance Systems (ADAS) applications has resulted in development of many new algorithms that are applied in solving various challenging problems. These algorithms need to be implemented on existing ADAS platforms which are usually heterogeneous in order to maximize computing power, while minimizing power consumption. The problem becomes how to efficiently decouple the algorithm and map parts of it to heterogeneous hardware, often including CPU, DSP and GPU blocks. This paper gives some insight into efficient ADAS algorithms mappings and optimizations for these platforms. As an illustrative example, driver monitoring algorithm is optimized.}, howpublished = {M33}, keywords = {ADAS, automotive, driver monitoring, ieeexplore}, pubstate = {published}, tppubtype = {conference} } Rapid expansion of Advanced Driver Assistance Systems (ADAS) applications has resulted in development of many new algorithms that are applied in solving various challenging problems. These algorithms need to be implemented on existing ADAS platforms which are usually heterogeneous in order to maximize computing power, while minimizing power consumption. The problem becomes how to efficiently decouple the algorithm and map parts of it to heterogeneous hardware, often including CPU, DSP and GPU blocks. This paper gives some insight into efficient ADAS algorithms mappings and optimizations for these platforms. As an illustrative example, driver monitoring algorithm is optimized. |