Autonomous Vehicle Software Symposium 2017 Image Map
 
Hotel:
Rai hotel logo

Official hotel agent!

From the publisher of:
Traffic Technology International Click here for free magazine subscription + media pack


Vision Zero International Click here for free magazine subscription + media pack


Crash Test Technology International Click here for free magazine subscription + media pack


 

2017 Conference Program

BUY YOUR DELEGATE PASS HERE

Day 1

Tuesday 20 June

09:00 - 12:30 - Keynote Presentations – Software Development & AI For Autonomous & Self-Driving Vehicles
Room C6

Moderator
Agustin Benito Bethencourt, principal consultant, Codethink, UK

09:00 - Selenium and Caesium – software elements of driverless cars
Paul Newman, founder, Oxbotica, UK
Selenium and Caesium are Oxbotica’s solution to this remarkable set of challenges. Selenium is our platform-agnostic autonomy operating system, which provides autonomy and deep scene understanding with any combination of laser or vision sensors, with or without prebuilt maps. Caesium is our fleet coordination system, which handles fleet configuration, data and learning sharing and meta information. This presentation will explain how these 'elements of autonomy' have been architected, built and deployed in fleets of vehicles and OEM vehicles.

09:30 - The open road to autonomous driving
Dan Cauchy, general manager automotive, The Linux Foundation, USA
The autonomous driving market continues to heat up as auto makers race to acquire tech companies or forge partnerships with suppliers who can accelerate and expand their development efforts. This flurry of M&A activity over the past year is being driven by the realisation that the software and hardware required for autonomous driving is too complex for any company to develop alone. Although this is a step in the right direction, an open-source platform for auto makers to share information such as map data, miles driven and scenarios tested could speed up production cycles and decrease time to market. This presentation will discuss the impact that open source could have on autonomous driving, and how collaboration could benefit auto makers without hindering competition.

10:00 - AI - The Answer to Autonomous Driving and Transportation
Serkan Arslan, Director of Automotive, Nvidia, GERMANY
Artificial Intelligence: What seemed like science fiction just a decade ago is now just science. In the next few years, AI will transform every major industry pertaining to the advancement of humankind. Soon, autonomous cars will reduce congestion and improve road safety. And the tools to embrace these improvements will be new infrastructures and vehicle computing platforms leveraging sensors such as cameras, radars, and lidars. Most importantly, this suite of sensors and tools includes AI for various levels of autonomous driving and transportation both inside and outside of cities. Discussed will be how AI algorithms leveraging structure from motion, sensor fusion and deep learning will help perceive the environment, create HD maps, and not only predict traffic behavior but what to do to control it.

10:30 - 11:00 - Break

11:00 - Collaborative Development for Basic Building Blocks of Autonomous Driving Systems
Lukas Bulwahn, software engineer - software infrastructure group, BMW Car IT, GERMANY
Since 2013, Tilmann Ochs, Daniel Wagner and Lukas Bulwahn have been working on research activities to define, motivate and implement a software platform for autonomous driving systems using custom-off-the-shelf open-source software. In 2014 they presented their understanding of future automotive software and their plans to use pre-existing open-source software for implementation of a collaborative automotive base platform at various automotive software engineering conferences. Now, three years later, it is time to re-evaluate this effort and critically review its progress, its successes and its failures. In this talk, the engineers present their main assumptions in 2013, and give some insights into ongoing software development activities supporting their ideas. Then, they evaluate to what extent they were successful in implementing these ideas, to what extent they could improve their understanding, and how this has refined their plans. On the technical side, they discuss the factors that influence the selection of the communication middleware and the underlying operating system of an automotive platform for autonomous driving systems. On the business side, they discuss the economics of automotive software development and the implications for the use and development of open-source software in the automotive domain.

11:30 - The ecosystem of self-driving by AImotive
Árpád Takács, outreach scientist, AImotive, HUNGARY
AImotive has developed a full-stack software suite for fully autonomous self-driving cars, providing a hardware-agnostic, scalable solution. Based on the idea that self-driving cars should mimic human behaviour, the algorithms rely on cameras as primary sensors for accomplishing the tasks of object recognition and classification, localisation, decision making, trajectory planning and vehicle control. The software engine components are aided by an extensive toolkit to accelerate the training and verification, including calibration, data collection and augmented data generation, semi-supervised annotation, and a real-time, photorealistic simulation environment. The presentation will provide an insight into the ecosystem of these software components.

12:00 - 12:30 - Panel Discussion Topic 1 - Collaboration, regulation and legislation. Topic 2 - AI and algorithms.
The panel will investigate if open source and databases can help software engineers limit liability for OEMS, as well as addressing the challenges of integrating AI into self-driving algorithms.
Paul Newman, founder, Oxbotica, UK
Dan Cauchy, general manager automotive, The Linux Foundation, USA
Serkan Arslan, Director of Automotive, Nvidia, GERMANY
Lukas Bulwahn, software engineer - software infrastructure group, BMW Car IT, GERMANY
Árpád Takács, outreach scientist, AImotive, HUNGARY


Moderator:
Agustin Benito Bethencourt, principal consultant , Codethink

12:30 - 14:00 - Lunch

14:00 - 17:30 - Solving Complex Software Challenges for Autonomous Transportation
Room C6

Moderator
Andrea Leitner, project manager and research engineer, AVL List GmbH, AUSTRIA

14:00 - Compete and succeed with IIoT architectures in autonomous cars
Bob Leigh, director of market development, Real-Time Innovations, USA
Dr Alexander Leonhardi, senior manager, ETAS, GERMANY
Autonomous vehicles are quickly becoming reality. Unfortunately, the existing industry standards can’t keep pace. New players and standards are driving OEMs, Tier 1s and electric car makers to adopt unfamiliar technologies and approaches. In particular, the DDS (Data Distribution Service) standard specifically addresses autonomy in other industries. It now presents an alternative architecture and communication framework for complex, safety-critical automotive designs. This session will review how DDS can solve the most challenging use cases in new car architectures. It will also outline how this IIoT standard will complement or impact AUTOSAR standards.

14:30 - A multi-platform development environment and core for autonomous vehicle software
Dr Joonwoo Son, Principal research Engineer, DGIST, KOREA
Benny Sjöstrand, CTO, Zuragon, SWEDEN
This presentation aims to give an overview of how to use ViCANdo as a development platform for autonomous vehicle software developers, and the sensor fusion connected to this task. The ViCANdo development platform is a cross-platform development system (Win, IoS, Android, QNX, Greenhills) utilising Matlabs/Simulink OpenGL, and OpenCV for vehicle, pedestrian and road sign detection and compiles to various target environments. A complete set up and build environment for an ADAS app for an Android target will be demonstrated and discussed during the presentation.

15:00 - The open standards enabling vision processing in ADAS
Illya Rudkin, principal software engineer, Codeplay Ltd, UK
For cars to control or drive themselves, they must 'see' their surroundings in a fast, safe and secure way. Recent advances in computing hardware make vision algorithms and machine learning practical for automotive systems. Open software standards allow competition and innovation in a supply chain, providing substantial advantages to all involved. Open standards enable innovators to develop software for vision or deep learning with flexibility of different hardware. Codeplay will present the virtues of open software standards OpenCL and its growing use in automotive, plus how its implementation stack includes open standard SYCL and Google Tensaflow to enable machine learning solutions.

15:30 - 16:00 - Break

16:00 - Supporting standard CNN networks on manycore processors
Dr Benoît Dupont de Dinechin, CTO, Kalray, FRANCE
KaNN is a domain-specific code generator for CNN inference that targets manycore processors featuring software programmable cores, local and DDR memories. Input is a standard CNN description such as Berkeley Caffe prototxt file, augmented with the parameters obtained during training. The KaNN generated code is optimized for low-latency parallel execution and effective DDR bandwidth exploitation. This result is obtained through the exploration and the selection of one of the many compute graph that represent the CNN forward computation, and the evaluation of this graph across the DDR and local memories according to a macro-pipelining scheme. The KaNN compute graphs have nodes that correspond either to 3D tiles of the images that represent the results of layer processing, or to operators on these tiles including 1x1 convolution kernels, pointwise accumulation and concatenation. The generated code is a collection of C kernels with asynchronous DMA transfers between the local and DDR memories. Kalray will be presenting KaNN and how it has been built to execute efficient CNN.

16:30 - The pedestrian challenge
Michael Hartmann, researcher, Virtual Vehicle Research Center, AUSTRIA
When looking at 'how to make future automated vehicles save', the topic of trajectory planning plays a key role. How can safe trajectories be found for autonomous vehicles? Motion planning for automated vehicles moving in traffic situations in general requires decision making based on incomplete and uncertain knowledge of the situation. The importance of trajectory planning is most evident for situations with vulnerable road users. What happens if some assumptions for planning are uncertain? It is proposed to use methods for uncertainty quantification in motion planning considering the fact that the environment can't be modelled.

17:00 - Addressing software complexity in ADAS and autonomous driving
Christopher Wild, technical director - CoherenSE, Altran, FRANCE
The future of automotive is increasingly defined by intelligent software. Such software will support increasingly sophisticated algorithms and services for ADAS, with the goal of widespread adoption of autonomous driving. Creating the required software, integrating and testing it, and managing it in-field present a challenge of complexity. Addressing this complexity requires a change in culture and architecture in the automotive world. In this presentation, we introduce CoherenSE and VueForge for ADAS verification, which provide an approach to software architecture targeting the management of complexity in development, integration and test of ADAS.


Day 2

Wednesday 21 June

09:00 - 13:00 - Solving Complex Software Challenges for Autonomous Transportation - Continued
Room C6

Moderator
Bob Leigh, director of market development, Real-Time Innovations, USA

09:00 - Collaborative software development and software aggregation for autonomous vehicle control systems
Randolph Toom, director innovation, Faar Industry, FRANCE
FAAR Industry has brought its Self-e-Car open development platform to a higher level of demonstration by integrating hardware and software technology bricks that originate from its own engineering department and several research institutes and suppliers of sensors. Fully operational sensor fusion with radar, lidar and camera, mapping procedures, trip planning and vehicle control strategies have been merged into one executable and flashed onto a single ECU. The challenge in this process was to structure the global SW and each brick to enable plug and play of SW bricks from different sources, and to fit everything in only one embedded ECU.

09:30 - Making future autonomous transportation a reality – safe and secure
John Round, Director of Software Engineering, NXP Semiconductors, USA
As the world’s largest supplier of automotive semiconductor solutions, NXP is driving innovation for safe and secure vehicle autonomy. In this presentation NXP will discuss the prevailing software and systems architectures that are evolving to address the challenges of fully autonomous design. In addition, NXP will review emerging software technologies in areas such as tooling, AI, middleware and testing, and their potential contribution to mass-market deployment of safe autonomous vehicles. NXP will also share its unique perspective on the challenges and opportunities that exist to truly deliver our collective safe, secure and autonomous transportation future.

10:00 - Revision conformity, corporate governance and compliance in software calibration processes
Thomas Wambera, affiliate business manager, AVL Deutschland GmbH, GERMANY
The complete documentation of software application responsibilities, calibration processes and related data, as well as the possibility to evaluate these according to different aspects in terms of the right-to-the-right, is crucial when developing ADAS systems. By combining established and controlled workflows and measures for IT security, revision certainty and conformity with ISO 27001, European law (eIDAS regulation) and usability for the implementation of ISO 26262 can be achieved. The speech gives an introduction to legal and technical requirements for the documentation, and is intended to provide an awareness of the complexity and dependencies in the software application process.

10:30 - 11:00 - Break

Moderator
John Round, Director of Software Engineering, NXP Semiconductors, USA

11:00 - Mastering time coherency and execution performance in your AD software
Nicolas Du Lac, CEO, Intempora SA, FRANCE
As embedded software in intelligent vehicles becomes more and more complex, because the number of ECUs and parallel software tasks has increased tremendously, it becomes critical to set up mechanisms that can handle time coherency among these software tasks and data streams. In this presentation we will address some of the software design concepts in the RTMaps framework that allow time coherency to be mastered in multi-thread and distributed architectures while achieving unprecedented execution performance in applications for autonomous driving, from the early stages of applications prototyping down to the execution on the most recent ECU architectures.

11:30 - The nervous system of an automated driving vehicle
Peter Brink, principal engineer, PolySync Technologies Inc, USA
There are many discussions about the 'brain' of an automated driving vehicle (ADV), looking at the technology required to perform decision making in real time. Little has been said about the nervous system of the ADV – the mechanisms by which the environmental data around the vehicle is transmitted to and from the brain. This presentation covers the requirements for a vehicle hardware abstraction layer for the software defined vehicle. The system requirements end up making the environmental data one of the most safety-critical pieces. How that data is shared is a key factor in making this abstraction layer a requirement.

12:00 - Towards autonomous driving – developing reliable automated driving features
Diego Barral, Senior Application Engineer, MathWorks, GERMANY
Building autonomous cars requires overcoming several technological challenges. It mainly contains three elements: perception, planning and control. This paper will demonstrate how to use a model-based engineering workflow solution based on MATLAB and Simulink to develop and efficiently test automated driving features including different sensor technology such as camera, radar and lidar.

12:30 - Highly automated driving on highways – reference architecture for coping with complexity
Sebastian Klaas, senior project manager, Elektrobit, GERMANY
The complexity of automated driving systems (e.g. adaptive cruise control, lane keeping assist, emergency braking, etc.) and related components has been rapidly increasing during recent years. Targeting Level 3 and above functions, like a Highway Pilot, adds another level of complexity due to the necessity for cooperation between formerly independent functionalities and additional safety needs. This talk discusses how to cope with complexity, based on a reference architecture and implementation for a Highway Pilot system.

13:00 - 13:30 - Lunch

14:30 - 17:30 - Building Resilient Software-Based Security Systems for Autonomous Vehicles
Room C6

Moderator
Dr Mehrdad Dianati, associate professor, University of Surrey, UK

14:30 - Protecting automotive sensors in connected cars: emerging trends with security vulnerabilities and software solutions
Frank Sattler, Director of Business Development, Automotive, EMEA, Irdeto, GERMANY
The increased connectivity and complexity in modern vehicles is resulting in new risks and threats. Hackers continuously evolve attack strategies to exploit vulnerabilities and access vehicles. To address these challenges, the industry must make cybersecurity a priority by focusing on key vulnerabilities that hackers exploit. Automotive sensors' designs – the embedded computers of which they are increasingly comprised – will be part of this software solution in concert with other systems to protect connected vehicles. This session will provide an overview of current threats, hackers' motivations for attacking automotive sensors, and software security strategies to protect automotive sensors.

15:00 - Lessons from the spaceship for the sedan
Frédéric Bourcier, delivery director, autonomous driving solutions, Wind River, FRANCE
As autonomous driving prototypes are being rolled out globally, the industry is developing in earnest. Safety is central. The auto industry can learn from other industries, such as aerospace and defence, with similarly tight restrictions. The industry is experiencing a wave of new innovation, and software is at the heart of this transformation. However, the development of these new autonomous systems requires a combination of mature and sound approaches as well as new ones, in order to ensure safe solutions. This presentation will cover: parallels between automotive and other mission-critical industries as they pertain to safety and security; the changing software landscape, especially with the growing trend of IoT, and how it is adding complexity to automotive systems; how autonomous vehicles can/will need to securely 'talk' with their surroundings in the smart cities of the future.

15:30 - 16:00 - Break

16:00 - How to safeguard software in vehicles
Dr Erik Oltmans, CEO, Software Improvement Group, NETHERLANDS
Connected cars represent a huge expansion in functionality, but one major question looms over the automotive ecosystem: How can we trust the underlying software? Using lessons learned the hard way in enterprise environments, Erik Oltmans, CEO of Software Improvement Group (SIG), will address gaps in software quality and how to improve the stability and security of systems both in and outside the vehicle—and explain why it all begins with early code inspection. Given the huge costs associated with recalls and other brand-damaging incidents—not to mention the impacts on human safety—the industry cannot afford to get this wrong. Oltmans will demonstrate how to get it right.

16:30 - Sealing connected and autonomous cars’ ECUs according to factory settings, by integrating with its software build environment
Pavel Zhytko, Full stack developer and web solutions architect, Karamba Security Ltd, BELARUS
The presentation will discuss how to automatically lock down the ECU according to its factory settings, by identification and mapping of all legitimate binaries and valid function calls. It will also cover checking all operations in runtime, blocking droppers and in-memory attacks as they don't comply with the factory settings, plus preventing cyber-attacks with zero false positives and negligible performance impact.

17:00 - The importance of development and delivery process transformation in the software supply chain for safety-critical environments when introducing upstream (open source) software
Agustin Benito Bethencourt, principal consultant, Codethink, UK
As more and more automotive systems and applications are built leveraging open-source software developed upstream, the automotive industry needs to adopt entirely new processes to adjust to consuming software from sources outside the traditional software supply chain. Particularly when it comes to safety-critical environments, transformations need to take place in existing development and delivery processes to ensure that delivered software can be trusted, with concepts such as chain of custody, provenance and reproducibility at the heart of this change.

*This Program may be subject to change.

Day 1

Tuesday 20 June

09:00 - 12:30 - Keynote Presentations – Software Development & AI For Autonomous & Self-Driving Vehicles
Room C6

Moderator
Agustin Benito Bethencourt, principal consultant, Codethink, UK

09:00 - Selenium and Caesium – software elements of driverless cars
Paul Newman, founder, Oxbotica, UK
Selenium and Caesium are Oxbotica’s solution to this remarkable set of challenges. Selenium is our platform-agnostic autonomy operating system, which provides autonomy and deep scene understanding with any combination of laser or vision sensors, with or without prebuilt maps. Caesium is our fleet coordination system, which handles fleet configuration, data and learning sharing and meta information. This presentation will explain how these 'elements of autonomy' have been architected, built and deployed in fleets of vehicles and OEM vehicles.

09:30 - The open road to autonomous driving
Dan Cauchy, general manager automotive, The Linux Foundation, USA
The autonomous driving market continues to heat up as auto makers race to acquire tech companies or forge partnerships with suppliers who can accelerate and expand their development efforts. This flurry of M&A activity over the past year is being driven by the realisation that the software and hardware required for autonomous driving is too complex for any company to develop alone. Although this is a step in the right direction, an open-source platform for auto makers to share information such as map data, miles driven and scenarios tested could speed up production cycles and decrease time to market. This presentation will discuss the impact that open source could have on autonomous driving, and how collaboration could benefit auto makers without hindering competition.

10:00 - AI - The Answer to Autonomous Driving and Transportation
Serkan Arslan, Director of Automotive, Nvidia, GERMANY
Artificial Intelligence: What seemed like science fiction just a decade ago is now just science. In the next few years, AI will transform every major industry pertaining to the advancement of humankind. Soon, autonomous cars will reduce congestion and improve road safety. And the tools to embrace these improvements will be new infrastructures and vehicle computing platforms leveraging sensors such as cameras, radars, and lidars. Most importantly, this suite of sensors and tools includes AI for various levels of autonomous driving and transportation both inside and outside of cities. Discussed will be how AI algorithms leveraging structure from motion, sensor fusion and deep learning will help perceive the environment, create HD maps, and not only predict traffic behavior but what to do to control it.

10:30 - 11:00 - Break

11:00 - Collaborative Development for Basic Building Blocks of Autonomous Driving Systems
Lukas Bulwahn, software engineer - software infrastructure group, BMW Car IT, GERMANY
Since 2013, Tilmann Ochs, Daniel Wagner and Lukas Bulwahn have been working on research activities to define, motivate and implement a software platform for autonomous driving systems using custom-off-the-shelf open-source software. In 2014 they presented their understanding of future automotive software and their plans to use pre-existing open-source software for implementation of a collaborative automotive base platform at various automotive software engineering conferences. Now, three years later, it is time to re-evaluate this effort and critically review its progress, its successes and its failures. In this talk, the engineers present their main assumptions in 2013, and give some insights into ongoing software development activities supporting their ideas. Then, they evaluate to what extent they were successful in implementing these ideas, to what extent they could improve their understanding, and how this has refined their plans. On the technical side, they discuss the factors that influence the selection of the communication middleware and the underlying operating system of an automotive platform for autonomous driving systems. On the business side, they discuss the economics of automotive software development and the implications for the use and development of open-source software in the automotive domain.

11:30 - The ecosystem of self-driving by AImotive
Árpád Takács, outreach scientist, AImotive, HUNGARY
AImotive has developed a full-stack software suite for fully autonomous self-driving cars, providing a hardware-agnostic, scalable solution. Based on the idea that self-driving cars should mimic human behaviour, the algorithms rely on cameras as primary sensors for accomplishing the tasks of object recognition and classification, localisation, decision making, trajectory planning and vehicle control. The software engine components are aided by an extensive toolkit to accelerate the training and verification, including calibration, data collection and augmented data generation, semi-supervised annotation, and a real-time, photorealistic simulation environment. The presentation will provide an insight into the ecosystem of these software components.

12:00 - 12:30 - Panel Discussion Topic 1 - Collaboration, regulation and legislation. Topic 2 - AI and algorithms.
The panel will investigate if open source and databases can help software engineers limit liability for OEMS, as well as addressing the challenges of integrating AI into self-driving algorithms.
Paul Newman, founder, Oxbotica, UK
Dan Cauchy, general manager automotive, The Linux Foundation, USA
Serkan Arslan, Director of Automotive, Nvidia, GERMANY
Lukas Bulwahn, software engineer - software infrastructure group, BMW Car IT, GERMANY
Árpád Takács, outreach scientist, AImotive, HUNGARY


Moderator:
Agustin Benito Bethencourt, principal consultant , Codethink

12:30 - 14:00 - Lunch

14:00 - 17:30 - Solving Complex Software Challenges for Autonomous Transportation
Room C6

Moderator
Andrea Leitner, project manager and research engineer, AVL List GmbH, AUSTRIA

14:00 - Compete and succeed with IIoT architectures in autonomous cars
Bob Leigh, director of market development, Real-Time Innovations, USA
Dr Alexander Leonhardi, senior manager, ETAS, GERMANY
Autonomous vehicles are quickly becoming reality. Unfortunately, the existing industry standards can’t keep pace. New players and standards are driving OEMs, Tier 1s and electric car makers to adopt unfamiliar technologies and approaches. In particular, the DDS (Data Distribution Service) standard specifically addresses autonomy in other industries. It now presents an alternative architecture and communication framework for complex, safety-critical automotive designs. This session will review how DDS can solve the most challenging use cases in new car architectures. It will also outline how this IIoT standard will complement or impact AUTOSAR standards.

14:30 - A multi-platform development environment and core for autonomous vehicle software
Dr Joonwoo Son, Principal research Engineer, DGIST, KOREA
Benny Sjöstrand, CTO, Zuragon, SWEDEN
This presentation aims to give an overview of how to use ViCANdo as a development platform for autonomous vehicle software developers, and the sensor fusion connected to this task. The ViCANdo development platform is a cross-platform development system (Win, IoS, Android, QNX, Greenhills) utilising Matlabs/Simulink OpenGL, and OpenCV for vehicle, pedestrian and road sign detection and compiles to various target environments. A complete set up and build environment for an ADAS app for an Android target will be demonstrated and discussed during the presentation.

15:00 - The open standards enabling vision processing in ADAS
Illya Rudkin, principal software engineer, Codeplay Ltd, UK
For cars to control or drive themselves, they must 'see' their surroundings in a fast, safe and secure way. Recent advances in computing hardware make vision algorithms and machine learning practical for automotive systems. Open software standards allow competition and innovation in a supply chain, providing substantial advantages to all involved. Open standards enable innovators to develop software for vision or deep learning with flexibility of different hardware. Codeplay will present the virtues of open software standards OpenCL and its growing use in automotive, plus how its implementation stack includes open standard SYCL and Google Tensaflow to enable machine learning solutions.

15:30 - 16:00 - Break

16:00 - Supporting standard CNN networks on manycore processors
Dr Benoît Dupont de Dinechin, CTO, Kalray, FRANCE
KaNN is a domain-specific code generator for CNN inference that targets manycore processors featuring software programmable cores, local and DDR memories. Input is a standard CNN description such as Berkeley Caffe prototxt file, augmented with the parameters obtained during training. The KaNN generated code is optimized for low-latency parallel execution and effective DDR bandwidth exploitation. This result is obtained through the exploration and the selection of one of the many compute graph that represent the CNN forward computation, and the evaluation of this graph across the DDR and local memories according to a macro-pipelining scheme. The KaNN compute graphs have nodes that correspond either to 3D tiles of the images that represent the results of layer processing, or to operators on these tiles including 1x1 convolution kernels, pointwise accumulation and concatenation. The generated code is a collection of C kernels with asynchronous DMA transfers between the local and DDR memories. Kalray will be presenting KaNN and how it has been built to execute efficient CNN.

16:30 - The pedestrian challenge
Michael Hartmann, researcher, Virtual Vehicle Research Center, AUSTRIA
When looking at 'how to make future automated vehicles save', the topic of trajectory planning plays a key role. How can safe trajectories be found for autonomous vehicles? Motion planning for automated vehicles moving in traffic situations in general requires decision making based on incomplete and uncertain knowledge of the situation. The importance of trajectory planning is most evident for situations with vulnerable road users. What happens if some assumptions for planning are uncertain? It is proposed to use methods for uncertainty quantification in motion planning considering the fact that the environment can't be modelled.

17:00 - Addressing software complexity in ADAS and autonomous driving
Christopher Wild, technical director - CoherenSE, Altran, FRANCE
The future of automotive is increasingly defined by intelligent software. Such software will support increasingly sophisticated algorithms and services for ADAS, with the goal of widespread adoption of autonomous driving. Creating the required software, integrating and testing it, and managing it in-field present a challenge of complexity. Addressing this complexity requires a change in culture and architecture in the automotive world. In this presentation, we introduce CoherenSE and VueForge for ADAS verification, which provide an approach to software architecture targeting the management of complexity in development, integration and test of ADAS.

*This Program may be subject to change.

Day 2

Wednesday 21 June

09:00 - 13:00 - Solving Complex Software Challenges for Autonomous Transportation - Continued
Room C6

Moderator
Bob Leigh, director of market development, Real-Time Innovations, USA

09:00 - Collaborative software development and software aggregation for autonomous vehicle control systems
Randolph Toom, director innovation, Faar Industry, FRANCE
FAAR Industry has brought its Self-e-Car open development platform to a higher level of demonstration by integrating hardware and software technology bricks that originate from its own engineering department and several research institutes and suppliers of sensors. Fully operational sensor fusion with radar, lidar and camera, mapping procedures, trip planning and vehicle control strategies have been merged into one executable and flashed onto a single ECU. The challenge in this process was to structure the global SW and each brick to enable plug and play of SW bricks from different sources, and to fit everything in only one embedded ECU.

09:30 - Making future autonomous transportation a reality – safe and secure
John Round, Director of Software Engineering, NXP Semiconductors, USA
As the world’s largest supplier of automotive semiconductor solutions, NXP is driving innovation for safe and secure vehicle autonomy. In this presentation NXP will discuss the prevailing software and systems architectures that are evolving to address the challenges of fully autonomous design. In addition, NXP will review emerging software technologies in areas such as tooling, AI, middleware and testing, and their potential contribution to mass-market deployment of safe autonomous vehicles. NXP will also share its unique perspective on the challenges and opportunities that exist to truly deliver our collective safe, secure and autonomous transportation future.

10:00 - Revision conformity, corporate governance and compliance in software calibration processes
Thomas Wambera, affiliate business manager, AVL Deutschland GmbH, GERMANY
The complete documentation of software application responsibilities, calibration processes and related data, as well as the possibility to evaluate these according to different aspects in terms of the right-to-the-right, is crucial when developing ADAS systems. By combining established and controlled workflows and measures for IT security, revision certainty and conformity with ISO 27001, European law (eIDAS regulation) and usability for the implementation of ISO 26262 can be achieved. The speech gives an introduction to legal and technical requirements for the documentation, and is intended to provide an awareness of the complexity and dependencies in the software application process.

10:30 - 11:00 - Break

Moderator
John Round, Director of Software Engineering, NXP Semiconductors, USA

11:00 - Mastering time coherency and execution performance in your AD software
Nicolas Du Lac, CEO, Intempora SA, FRANCE
As embedded software in intelligent vehicles becomes more and more complex, because the number of ECUs and parallel software tasks has increased tremendously, it becomes critical to set up mechanisms that can handle time coherency among these software tasks and data streams. In this presentation we will address some of the software design concepts in the RTMaps framework that allow time coherency to be mastered in multi-thread and distributed architectures while achieving unprecedented execution performance in applications for autonomous driving, from the early stages of applications prototyping down to the execution on the most recent ECU architectures.

11:30 - The nervous system of an automated driving vehicle
Peter Brink, principal engineer, PolySync Technologies Inc, USA
There are many discussions about the 'brain' of an automated driving vehicle (ADV), looking at the technology required to perform decision making in real time. Little has been said about the nervous system of the ADV – the mechanisms by which the environmental data around the vehicle is transmitted to and from the brain. This presentation covers the requirements for a vehicle hardware abstraction layer for the software defined vehicle. The system requirements end up making the environmental data one of the most safety-critical pieces. How that data is shared is a key factor in making this abstraction layer a requirement.

12:00 - Towards autonomous driving – developing reliable automated driving features
Diego Barral, Senior Application Engineer, MathWorks, GERMANY
Building autonomous cars requires overcoming several technological challenges. It mainly contains three elements: perception, planning and control. This paper will demonstrate how to use a model-based engineering workflow solution based on MATLAB and Simulink to develop and efficiently test automated driving features including different sensor technology such as camera, radar and lidar.

12:30 - Highly automated driving on highways – reference architecture for coping with complexity
Sebastian Klaas, senior project manager, Elektrobit, GERMANY
The complexity of automated driving systems (e.g. adaptive cruise control, lane keeping assist, emergency braking, etc.) and related components has been rapidly increasing during recent years. Targeting Level 3 and above functions, like a Highway Pilot, adds another level of complexity due to the necessity for cooperation between formerly independent functionalities and additional safety needs. This talk discusses how to cope with complexity, based on a reference architecture and implementation for a Highway Pilot system.

13:00 - 13:30 - Lunch

14:30 - 17:30 - Building Resilient Software-Based Security Systems for Autonomous Vehicles
Room C6

Moderator
Dr Mehrdad Dianati, associate professor, University of Surrey, UK

14:30 - Protecting automotive sensors in connected cars: emerging trends with security vulnerabilities and software solutions
Frank Sattler, Director of Business Development, Automotive, EMEA, Irdeto, GERMANY
The increased connectivity and complexity in modern vehicles is resulting in new risks and threats. Hackers continuously evolve attack strategies to exploit vulnerabilities and access vehicles. To address these challenges, the industry must make cybersecurity a priority by focusing on key vulnerabilities that hackers exploit. Automotive sensors' designs – the embedded computers of which they are increasingly comprised – will be part of this software solution in concert with other systems to protect connected vehicles. This session will provide an overview of current threats, hackers' motivations for attacking automotive sensors, and software security strategies to protect automotive sensors.

15:00 - Lessons from the spaceship for the sedan
Frédéric Bourcier, delivery director, autonomous driving solutions, Wind River, FRANCE
As autonomous driving prototypes are being rolled out globally, the industry is developing in earnest. Safety is central. The auto industry can learn from other industries, such as aerospace and defence, with similarly tight restrictions. The industry is experiencing a wave of new innovation, and software is at the heart of this transformation. However, the development of these new autonomous systems requires a combination of mature and sound approaches as well as new ones, in order to ensure safe solutions. This presentation will cover: parallels between automotive and other mission-critical industries as they pertain to safety and security; the changing software landscape, especially with the growing trend of IoT, and how it is adding complexity to automotive systems; how autonomous vehicles can/will need to securely 'talk' with their surroundings in the smart cities of the future.

15:30 - 16:00 - Break

16:00 - How to safeguard software in vehicles
Dr Erik Oltmans, CEO, Software Improvement Group, NETHERLANDS
Connected cars represent a huge expansion in functionality, but one major question looms over the automotive ecosystem: How can we trust the underlying software? Using lessons learned the hard way in enterprise environments, Erik Oltmans, CEO of Software Improvement Group (SIG), will address gaps in software quality and how to improve the stability and security of systems both in and outside the vehicle—and explain why it all begins with early code inspection. Given the huge costs associated with recalls and other brand-damaging incidents—not to mention the impacts on human safety—the industry cannot afford to get this wrong. Oltmans will demonstrate how to get it right.

16:30 - Sealing connected and autonomous cars’ ECUs according to factory settings, by integrating with its software build environment
Pavel Zhytko, Full stack developer and web solutions architect, Karamba Security Ltd, BELARUS
The presentation will discuss how to automatically lock down the ECU according to its factory settings, by identification and mapping of all legitimate binaries and valid function calls. It will also cover checking all operations in runtime, blocking droppers and in-memory attacks as they don't comply with the factory settings, plus preventing cyber-attacks with zero false positives and negligible performance impact.

17:00 - The importance of development and delivery process transformation in the software supply chain for safety-critical environments when introducing upstream (open source) software
Agustin Benito Bethencourt, principal consultant, Codethink, UK
As more and more automotive systems and applications are built leveraging open-source software developed upstream, the automotive industry needs to adopt entirely new processes to adjust to consuming software from sources outside the traditional software supply chain. Particularly when it comes to safety-critical environments, transformations need to take place in existing development and delivery processes to ensure that delivered software can be trusted, with concepts such as chain of custody, provenance and reproducibility at the heart of this change.

*This Program may be subject to change.

 
 

Opening Networking Breakfast




Conference Programme



Preliminary Conference Programme Online Now

Click here to view the programme

 
Held alongside:
 
Co-Located with: