Autonomous Vehicle Test & Development Symposium 2017
 
Media patners:








 
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


Automotive Testing Technology International Click here for free magazine subscription + media pack


Automotive Testing Technology International Click here for free magazine subscription + media pack


Electric & Hybrid Vehicle Technology International Click here for free magazine subscription + media pack

 

2017 Conference Programme

BUY YOUR DELEGATE PASS HERE

Day 1

Wednesday 25 October

08:15 - 08:55 - Networking Breakfast

Join us on the opening morning for our complimentary networking breakfast. All speakers, delegates, and sponsors are invited to attend.

09:00 - 12:30 - Opening Session

Scenario generation for L4 automated vehicles
Phil Magney, principal advisor, Vision Systems Intelligence LLC, USA
When is a system 'good enough'? Which scenarios should be tested until a system can be released? How many kilometers need to be driven to prove that a system is robust and failsafe? And should there be a retest of all scenarios after a hardware or software update? When developing L4 applications it is extremely important to create a library of scenarios and models to validate features. Scenarios and scenario models contribute not only to performance validation but also to system requirements. Furthermore, several non-functional and performance requirements are driven from use cases, which bring significant value in autonomous vehicle development. In this session, VSI examines the latest approaches for scenario mining as well as best practices for converting those into viable simulations.

New testing requirements: potential roles of track, simulation and on-road testing
Dr Shawn Kimmel, lead technologist, Booz Allen Hamilton, USA
Testing of highly automated vehicles (HAVs) will require new tools and methods to address the increased complexity of these technologies. Numerous potential scenarios may be encountered by HAVs; there is a challenge to create objective, repeatable and transparent tests to bring safe products to market. This talk will explore potential new testing requirements based on an analysis of proposed HAV functions, operational design domains, and object and event detection and recognition. A framework will be presented on the roles of simulation and on-road testing to enable rapid and inexpensive testing of a wide range of conditions and software pathways.

Innovative concepts for safely testing autonomous vehicles
Dr Edward Kraft, Associate Executive Director for research, University of Tennessee Space Institute, USA
Integrating digital engineering approaches developed by DoD and NASA, a shift to an architecture-centric, model-based systems engineering/model-based engineering approach can: (1) Provide a quantifiable, incremental validation of safety and maintain constant digital connectivity between requirements, design, performance and safety requirements over the lifecycle; (2) Shift discovery of defects and emergent behaviors to the left in the development cycle, increasing safety and reducing time to market; (3) Support application of non-deterministic control and monitoring methods to ensure fail-safe operations during testing; (4) Apply deep-learning big data analytics to fleet operational performance to improve design, development and test processes.

Applying artificial intelligence to achieve safe Level 5 autonomy
Dominic Gallello, president and CEO, MSC Software, USA
We have a long journey ahead to move from current offline simulation methods to testing for cars driving completely without human intervention. There are still many challenges to be solved, much technology to be invented and many more orders of magnitude of simulations to be performed than what is currently being done today. The use of artificial intelligence is the key to making safe self-driving cars possible. In this session, Dominic Gallello will provide an overview of the basics of artificial intelligence and the development of perception and decision neural networks necessary to deliver autonomous driving with high confidence.

13:30 - 17:00 - Afternoon Session

Designing sensor algorithms for the automobile environment
Tony Gioutsos, director, TASS International, USA
The difference between an automobile environment and other environments when designing an algorithm is substantial. When driving a vehicle at a high rate of speed, any error can produce tragic results. Because automobiles are expected to survive 15 years in all kinds of conditions, it is basically impossible to design sensor algorithms that are tested for all kinds of scenarios that could be encountered. In this paper, we outline a generic approach to designing sensor algorithms that are robust to the real world.

VeHiL real-time operating autonomous vehicle testbed
Prof Abdel Mayyas, assistant professor, Arizona State University, USA
Although model predictive control offers great promise to solve many energy and reliability problems in modern vehicles, transitioning from high to full autonomous driving across different contexts requires new creative methods for testing and validation. This work proposes a new deployment of the real-time operating connected vehicle testbed for VeHiL.

Signalized intersections prevent traveling from point A to point B legally
Brian Ceccarelli, principal engineer, Talus Software PLLC, USA
The implementation of signalized intersections makes it impossible for AV manufacturers to meet the requirement that vehicles shall travel from point A to point B legally and safely. There are five major defects in the physics of traffic signals that make it impossible for an AV to legally and safely navigate an intersection in all cases: 1) The universal application of special-case-only kinematic equation; 2) The use of the wrong initial velocity; 3) The misapplication of stochastic methods to human and vehicle factors; 4) The assumption that analytic solutions are physical solutions; 5) The omission of yellow light tolerance computations.

Validating advanced driver assistance systems using HiL test benches
Sean Wyatt, senior project engineer / program manager, engineering test solutions, ETAS Inc, USA
Hardware-in-the-loop (HiL) based testing methods offer the great advantage of validating components and systems at an early stage of the development cycle, and they are established in the automotive industry. When validating advanced driver assistance systems using HiL test benches, engineers face different barriers and conceptual difficulties: how to pipe simulated signals into multiple sensors including radar, ultrasonic, video or lidar; how to combine classical physical simulations, e.g. vehicle dynamics, with sophisticated three-dimensional, GPU-based environmental simulations. We discuss two categories: 1) Hardware level: communications structure; 2) Software level: providing data the UuT expects.

Panel Discussion

Dr Shawn Kimmel, lead technologist, Booz Allen Hamilton, USA
Dr Edward Kraft, Associate Executive Director for research, University of Tennessee Space Institute, USA
Brian Ceccarelli, principal engineer, Talus Software PLLC, USA
Tony Gioutsos, director, TASS International, USA
Prof Abdel Mayyas, assistant professor, Arizona State University, USA

Day 2

Thursday 26 October

09:00 - 12:30 - Morning Session

The use of physical validation space for autonomous vehicles
Paul Krutko, president and CEO, Ann Arbor SPARK, USA
This session will highlight the methodology utilized by purpose-built testing and validation centers, and the benefits of using this approach. Connected, automated and autonomous vehicle technology will change the level of safety on our roads. This technology has now been deployed on public roads, but further validation is needed to perfect the technology. The United States has designated 10 sites as proving grounds, based on the concept that there needs to be more attention dedicated to these efforts. Software-based testing used today is an important part of the validation scheme, but requires physical deployments to further validate.

Automotive radar simulation for ADAS and autonomous driving
Dr Sandeep Sovani, director, global automotive industry, ANSYS, USA
Simulation is crucial in automotive radar development since it provides precise insights into real-life radar operation, at a fraction of the cost and time needed for physical field tests. High-fidelity physics-based simulation techniques such as an electromagnetic field solver and a shooting-bouncing rays solver are needed to provide results with real-life accuracy. Simulation of five aspects of radar development will be presented: (a) antenna design; (b) isolated radar simulation; (c) as-installed radar simulation; (d) in-environment radar simulation; (e) in-driving scenario radar simulation.

A safe path: watching an autonomous system in operation
Paul Perrone, CEO/founder, Perrone Robotics Inc, USA
Software that replaces a human in an autonomous vehicle must be complex. It should evolve and improve over time as more situations are encountered. This model conflicts with current safety-critical standards where a precise and complete assessment of simpler, unchanging code against formal requirements and testing is required. How to proceed? We believe that although core elements of the software must be tested and trusted, higher-level functions must also be interrogated during operation. Our watchdog model evaluates key parameters and behaviors to ensure correct operation. If malfunction is detected, the system fails safely via a separate mechanism.

Cybersecurity for autonomous vehicles: problems and solutions
Dr Jeremy Straub, assistant professor, North Dakota State University, USA
This presentation provides an overview of the cybersecurity issues that currently face vehicles with limited autonomy as well as issues that will become more pronounced at higher vehicle autonomy levels. These issues are considered at both the individual vehicle and system-of-vehicles level. The efficacy of current cybersecurity techniques for responding to these problems is considered. Gaps that are not filled by conventional techniques are identified. Potential solutions for filling these gaps are suggested and a discussion of their feasibility and development pathway is presented. The use of intrusion detection systems as a catch-all for unexpected attack types is also discussed.

Integrating autonomous vehicles requires high reliability of ADS systems
Ralph Buckingham, director - connected/autonomous technologies, Intertek Transportation Technologies, USA
The reliability and performance of components used to perform automated driving system functions has to be of the highest caliber to ensure the safe integration of autonomous vehicles with other road users. This presentation will cover current technologies being used to automate vehicles, known reliability/performance challenges, and testing processes that can be leveraged to improve reliability. Specific areas of focus will be radar/lidar/camera technologies, the current state of each technology as it pertains to design robustness/reliability/performance, and the testing methods that can be used to improve reliability: design validation/accelerated stress testing/failure analysis.

Taking autonomous testing to the virtual streets
Scott Harvey, co-founder, Civil Maps, USA
Testing autonomous vehicles is both resource and time intensive. Driverless vehicles must be prepared to travel in different environments, weather conditions and with an infinite combination of sensor technologies. Civil Maps, the Ford-funded cognition systems provider for autonomous vehicles, has created a solution to test autonomous cars in a virtual environment. Its Synthetics platform is a software offering that trains driverless cars in simulated, procedurally generated worlds.

13:30 - 17:00 - Afternoon Session

Virtual testing of traffic scenarios using real-world data
Kunal Patil, senior applications engineer, dSpace Inc, USA
This presentation will give an overview of how you can get the real world to your virtual test drives for testing of traffic scenarios. It will show how to use the dSPACE ASM tool chain to define road networks, and vehicle and traffic object movements for testing of autonomous driving features.

Deploying automated mobility services: revolution of the transportation ecosystem
Leemor Chandally, director of strategic partnerships, BestMile, USA
Although incredibly advanced, autonomous vehicles are not sufficient to offer a mobility service. Optimization has to take place at the fleet scale in order to create a coordinated and efficient mobility service. Like two pieces of the same puzzle, autonomous vehicles and fleet optimization platforms are complementary and form a sustainable, efficient and revolutionary autonomous transportation system. BestMile was the first company offering on-demand autonomous mobility services and is now enabling the first instances in which electric autonomous shuttles are circulating through city centers on pedestrian areas and open roads, and servicing the public on a daily basis.

In-vehicle measurement and calibration of ADAS ECUs
Koos Zwaanenburg, program manager, ETAS, USA
ADAS automate, adapt and enhance automobiles for safety and better driving. ADAS will become a component of autonomously driving vehicles. ADAS uses different sensors than those used in automotive controllers for powertrains and chassis control, e.g. cameras. Thus, ADAS calibration engineers will have to record video data in the test vehicle, and time-align it with data from traditional sensors as well as with data computed inside the ADAS controller. This presentation will show how to simultaneously measure and align real-time data generated by ADAS controllers with data from video cameras, and with continuous-time electric signals from other sensors.

Panel Discussion

Paul Krutko, president and CEO, Ann Arbor SPARK, USA
Scott Harvey, co-founder, Civil Maps, USA
Paul Perrone, CEO/founder, Perrone Robotics Inc, USA
Dr Sandeep Sovani, director, global automotive industry, ANSYS, USA

*This Programme may be subject to change.

Day 1

Wednesday 25 October

08:15 - 08:55 - Networking Breakfast

Join us on the opening morning for our complimentary networking breakfast. All speakers, delegates, and sponsors are invited to attend.

09:00 - 12:30 - Opening Session

Scenario generation for L4 automated vehicles
Phil Magney, principal advisor, Vision Systems Intelligence LLC, USA
When is a system 'good enough'? Which scenarios should be tested until a system can be released? How many kilometers need to be driven to prove that a system is robust and failsafe? And should there be a retest of all scenarios after a hardware or software update? When developing L4 applications it is extremely important to create a library of scenarios and models to validate features. Scenarios and scenario models contribute not only to performance validation but also to system requirements. Furthermore, several non-functional and performance requirements are driven from use cases, which bring significant value in autonomous vehicle development. In this session, VSI examines the latest approaches for scenario mining as well as best practices for converting those into viable simulations.

New testing requirements: potential roles of track, simulation and on-road testing
Dr Shawn Kimmel, lead technologist, Booz Allen Hamilton, USA
Testing of highly automated vehicles (HAVs) will require new tools and methods to address the increased complexity of these technologies. Numerous potential scenarios may be encountered by HAVs; there is a challenge to create objective, repeatable and transparent tests to bring safe products to market. This talk will explore potential new testing requirements based on an analysis of proposed HAV functions, operational design domains, and object and event detection and recognition. A framework will be presented on the roles of simulation and on-road testing to enable rapid and inexpensive testing of a wide range of conditions and software pathways.

Innovative concepts for safely testing autonomous vehicles
Dr Edward Kraft, Associate Executive Director for research, University of Tennessee Space Institute, USA
Integrating digital engineering approaches developed by DoD and NASA, a shift to an architecture-centric, model-based systems engineering/model-based engineering approach can: (1) Provide a quantifiable, incremental validation of safety and maintain constant digital connectivity between requirements, design, performance and safety requirements over the lifecycle; (2) Shift discovery of defects and emergent behaviors to the left in the development cycle, increasing safety and reducing time to market; (3) Support application of non-deterministic control and monitoring methods to ensure fail-safe operations during testing; (4) Apply deep-learning big data analytics to fleet operational performance to improve design, development and test processes.

Applying artificial intelligence to achieve safe Level 5 autonomy
Dominic Gallello, president and CEO, MSC Software, USA
We have a long journey ahead to move from current offline simulation methods to testing for cars driving completely without human intervention. There are still many challenges to be solved, much technology to be invented and many more orders of magnitude of simulations to be performed than what is currently being done today. The use of artificial intelligence is the key to making safe self-driving cars possible. In this session, Dominic Gallello will provide an overview of the basics of artificial intelligence and the development of perception and decision neural networks necessary to deliver autonomous driving with high confidence.

13:30 - 17:00 - Afternoon Session

Designing sensor algorithms for the automobile environment
Tony Gioutsos, director, TASS International, USA
The difference between an automobile environment and other environments when designing an algorithm is substantial. When driving a vehicle at a high rate of speed, any error can produce tragic results. Because automobiles are expected to survive 15 years in all kinds of conditions, it is basically impossible to design sensor algorithms that are tested for all kinds of scenarios that could be encountered. In this paper, we outline a generic approach to designing sensor algorithms that are robust to the real world.

VeHiL real-time operating autonomous vehicle testbed
Prof Abdel Mayyas, assistant professor, Arizona State University, USA
Although model predictive control offers great promise to solve many energy and reliability problems in modern vehicles, transitioning from high to full autonomous driving across different contexts requires new creative methods for testing and validation. This work proposes a new deployment of the real-time operating connected vehicle testbed for VeHiL.

Signalized intersections prevent traveling from point A to point B legally
Brian Ceccarelli, principal engineer, Talus Software PLLC, USA
The implementation of signalized intersections makes it impossible for AV manufacturers to meet the requirement that vehicles shall travel from point A to point B legally and safely. There are five major defects in the physics of traffic signals that make it impossible for an AV to legally and safely navigate an intersection in all cases: 1) The universal application of special-case-only kinematic equation; 2) The use of the wrong initial velocity; 3) The misapplication of stochastic methods to human and vehicle factors; 4) The assumption that analytic solutions are physical solutions; 5) The omission of yellow light tolerance computations.

Validating advanced driver assistance systems using HiL test benches
Sean Wyatt, senior project engineer / program manager, engineering test solutions, ETAS Inc, USA
Hardware-in-the-loop (HiL) based testing methods offer the great advantage of validating components and systems at an early stage of the development cycle, and they are established in the automotive industry. When validating advanced driver assistance systems using HiL test benches, engineers face different barriers and conceptual difficulties: how to pipe simulated signals into multiple sensors including radar, ultrasonic, video or lidar; how to combine classical physical simulations, e.g. vehicle dynamics, with sophisticated three-dimensional, GPU-based environmental simulations. We discuss two categories: 1) Hardware level: communications structure; 2) Software level: providing data the UuT expects.

Panel Discussion

Dr Shawn Kimmel, lead technologist, Booz Allen Hamilton, USA
Dr Edward Kraft, Associate Executive Director for research, University of Tennessee Space Institute, USA
Brian Ceccarelli, principal engineer, Talus Software PLLC, USA
Tony Gioutsos, director, TASS International, USA
Prof Abdel Mayyas, assistant professor, Arizona State University, USA

*This Programme may be subject to change.

Day 2

Thursday 26 October

09:00 - 12:30 - Morning Session

The use of physical validation space for autonomous vehicles
Paul Krutko, president and CEO, Ann Arbor SPARK, USA
This session will highlight the methodology utilized by purpose-built testing and validation centers, and the benefits of using this approach. Connected, automated and autonomous vehicle technology will change the level of safety on our roads. This technology has now been deployed on public roads, but further validation is needed to perfect the technology. The United States has designated 10 sites as proving grounds, based on the concept that there needs to be more attention dedicated to these efforts. Software-based testing used today is an important part of the validation scheme, but requires physical deployments to further validate.

Automotive radar simulation for ADAS and autonomous driving
Dr Sandeep Sovani, director, global automotive industry, ANSYS, USA
Simulation is crucial in automotive radar development since it provides precise insights into real-life radar operation, at a fraction of the cost and time needed for physical field tests. High-fidelity physics-based simulation techniques such as an electromagnetic field solver and a shooting-bouncing rays solver are needed to provide results with real-life accuracy. Simulation of five aspects of radar development will be presented: (a) antenna design; (b) isolated radar simulation; (c) as-installed radar simulation; (d) in-environment radar simulation; (e) in-driving scenario radar simulation.

A safe path: watching an autonomous system in operation
Paul Perrone, CEO/founder, Perrone Robotics Inc, USA
Software that replaces a human in an autonomous vehicle must be complex. It should evolve and improve over time as more situations are encountered. This model conflicts with current safety-critical standards where a precise and complete assessment of simpler, unchanging code against formal requirements and testing is required. How to proceed? We believe that although core elements of the software must be tested and trusted, higher-level functions must also be interrogated during operation. Our watchdog model evaluates key parameters and behaviors to ensure correct operation. If malfunction is detected, the system fails safely via a separate mechanism.

Cybersecurity for autonomous vehicles: problems and solutions
Dr Jeremy Straub, assistant professor, North Dakota State University, USA
This presentation provides an overview of the cybersecurity issues that currently face vehicles with limited autonomy as well as issues that will become more pronounced at higher vehicle autonomy levels. These issues are considered at both the individual vehicle and system-of-vehicles level. The efficacy of current cybersecurity techniques for responding to these problems is considered. Gaps that are not filled by conventional techniques are identified. Potential solutions for filling these gaps are suggested and a discussion of their feasibility and development pathway is presented. The use of intrusion detection systems as a catch-all for unexpected attack types is also discussed.

Integrating autonomous vehicles requires high reliability of ADS systems
Ralph Buckingham, director - connected/autonomous technologies, Intertek Transportation Technologies, USA
The reliability and performance of components used to perform automated driving system functions has to be of the highest caliber to ensure the safe integration of autonomous vehicles with other road users. This presentation will cover current technologies being used to automate vehicles, known reliability/performance challenges, and testing processes that can be leveraged to improve reliability. Specific areas of focus will be radar/lidar/camera technologies, the current state of each technology as it pertains to design robustness/reliability/performance, and the testing methods that can be used to improve reliability: design validation/accelerated stress testing/failure analysis.

Taking autonomous testing to the virtual streets
Scott Harvey, co-founder, Civil Maps, USA
Testing autonomous vehicles is both resource and time intensive. Driverless vehicles must be prepared to travel in different environments, weather conditions and with an infinite combination of sensor technologies. Civil Maps, the Ford-funded cognition systems provider for autonomous vehicles, has created a solution to test autonomous cars in a virtual environment. Its Synthetics platform is a software offering that trains driverless cars in simulated, procedurally generated worlds.

13:30 - 17:00 - Afternoon Session

Virtual testing of traffic scenarios using real-world data
Kunal Patil, senior applications engineer, dSpace Inc, USA
This presentation will give an overview of how you can get the real world to your virtual test drives for testing of traffic scenarios. It will show how to use the dSPACE ASM tool chain to define road networks, and vehicle and traffic object movements for testing of autonomous driving features.

Deploying automated mobility services: revolution of the transportation ecosystem
Leemor Chandally, director of strategic partnerships, BestMile, USA
Although incredibly advanced, autonomous vehicles are not sufficient to offer a mobility service. Optimization has to take place at the fleet scale in order to create a coordinated and efficient mobility service. Like two pieces of the same puzzle, autonomous vehicles and fleet optimization platforms are complementary and form a sustainable, efficient and revolutionary autonomous transportation system. BestMile was the first company offering on-demand autonomous mobility services and is now enabling the first instances in which electric autonomous shuttles are circulating through city centers on pedestrian areas and open roads, and servicing the public on a daily basis.

In-vehicle measurement and calibration of ADAS ECUs
Koos Zwaanenburg, program manager, ETAS, USA
ADAS automate, adapt and enhance automobiles for safety and better driving. ADAS will become a component of autonomously driving vehicles. ADAS uses different sensors than those used in automotive controllers for powertrains and chassis control, e.g. cameras. Thus, ADAS calibration engineers will have to record video data in the test vehicle, and time-align it with data from traditional sensors as well as with data computed inside the ADAS controller. This presentation will show how to simultaneously measure and align real-time data generated by ADAS controllers with data from video cameras, and with continuous-time electric signals from other sensors.

Panel Discussion

Paul Krutko, president and CEO, Ann Arbor SPARK, USA
Scott Harvey, co-founder, Civil Maps, USA
Paul Perrone, CEO/founder, Perrone Robotics Inc, USA
Dr Sandeep Sovani, director, global automotive industry, ANSYS, USA

*This Programme may be subject to change.

 
 

15% Discount for USDOT and Federal Transportation Administrations

-

Click here!


Co-Located with:



 

Preliminary Conference Programme
now online!

click here!
Held alongside:

 
Topics under discussion:
  • Public road testing
  • Virtual testing
  • Simulation
  • Traffic scenario testing
  • Embedded software testing
  • Reliability testing of software and hardware systems
  • Safety and crash testing
  • Fail-safe testing
  • Cyber threat testing
  • Validation and verification
  • Autonomy software
  • VeHIL
  • V2V and V2X testing
  • Robotics
  • Testing legislation
  • Safety standards and legislation
  • Human factors and HMI testing
  • Case studies
  • Possibilities
  • Best practices