Autonomous Vehicle Test & Development Symposium 2017
Companies that have participated previously include
 
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

 
Hotel
Rai Hotel logo

Official hotel agent

 
From the organiser of:

Automotive Testing Expo Europe

 

2017 Conference Programme

BUY YOUR DELEGATE PASS HERE

Day 1

Tuesday 20 June

Keynote Presentations
Room A

Level 2 automation: lessons learned from crash investigation
Dr Ensar Becic, highway accident investigator / project manager, National Transportation Safety Board, USA
The first fatal crash involving an autonomous vehicle in the United States occurred in May 2016. The National Transportation Safety Board investigated this crash, in which a 2015 Tesla, operating in autopilot mode at that time, was travelling at 74mph on a state freeway, and struck and passed beneath the semitrailer of a combination vehicle that was completing a left turn. The autopilot did not detect the combination vehicle and the driver did not make any evasive manoeuvres. The investigation focused on (a) autopilot functionality and limitations, (b) driver trust and system misuse, (c) usability and standardisation of data.

Approaches for testing and type approval of automatic driving vehicles
Dr Houssem Abdellatif, global head autonomous driving & ADAS, TÜV SÜD Auto Service GmbH, GERMANY
Further road safety improvement is achievable through advanced driver assistance systems (ADAS) and automated driving (AD), which will result in autonomous driving in the future. However, common type approval procedures and testing approaches must be defined. The testing procedures based on physical testing seem to be insufficient to cover all the possible cases and thus to evaluate their efficiency and safety. Simulation methods seem to be the only feasible way to cover the huge number of possible scenarios. Physical tests will still be needed for verification and validation of these simulation models and setups.

Dynamic map layer technologies for intelligent automotive map data updates
Mike Tzamaloukas, vice president, autonomous drive and ADAS, and vice president, CoC navigation, Harman International Industries, Incorporated, GERMANY
As the number of sensors in cars increases, a technological solution for intelligent automotive map data updates can be used to create updates to existing road maps, and even support the necessary high-definition map content required for various safety and autonomy applications within the connected car. The presentation about dynamic map layer technologies will show how sensor fusion technology can be used to analyse safety and camera data to intelligently update map data.

Autonomous driving and dangerous weather-related road conditions?
Eric-Mark Huitema, global GM ground transport, IBM, The Weather Company, NETHERLANDS
An autonomously driving car cannot rely on the driver's interaction with the weather in its surroundings. However, weather is one of the main reasons for accidents and fatalities on safe roads. The solution must be a real-time and hyper-local weather interaction towards the car, so that the autonomous car can anticipate black ice, aquaplaning, thunderstorms, heavy rain or snow. This presentation will show how OEMs and aftermarket suppliers work with The Weather Company to make driverless interaction with the autonomous car possible.

Comprehensive modular verification and validation framework for automated systems
Andrea Leitner, project manager and research engineer, AVL List GmbH, AUSTRIA
This presentation aims to give an overview of the activities of ENABLE-S3, a large European project with around 71 partners from industry and research. The project’s goal is the provision of a comprehensive modular verification and validation framework to enable more efficient testing of automated systems. The framework does not only cover the testing platform, but also the testing methodology in order to reduce the number of tests. Based on requirements from 12 use cases in various domains, the result will be an approach to efficient and seamless scenario-based verification and validation in virtual, semi-virtual and real testing environments.

Open-Road and Real-World Testing
Room A

Moderator
Chris Reeves, Head of Connected & Autonmous Vehicles, Horiba Mira Ltd, UK

Navya worldwide development on private sites and open road
Christophe Sapet, CEO, Navya, FRANCE
The presentation will discuss autonomous vehicles at the heart of the smart city, and the benefits of autonomous vehicles on the first and last mile.

Development and testing of autonomous vehicles in urban environments
Dr Byoung-Kwang Kim, senior research engineer, Hyundai Motor Company, KOREA
We developed three autonomous driving IONIQs and two autonomous driving Tucson fuel cell electric vehicles. Users can call the vehicle like a taxi using smartphones and go to the destination autonomously. We tested them inside Hyundai’s Research and Development campus in Namyang, South Korea. Driving in the campus is a challenge because of the more than 10,000 employees on-site every day and almost urban environments. More than 1,000 volunteers, who are Hyundai's R&D engineers working in various divisions, experienced the full scenario of our autonomous vehicles. They assessed autonomous driving function and HMI.

GATEway – results from testing automated vehicles in Greenwich, London
Prof Nick Reed, academy director, TRL, UK
In this session, results from the GATEway project will be presented. This includes the testing of automated shuttle vehicles, a simulator study examining driver behaviour in the presence of automated vehicles, and the creation of a teleoperated vehicle capable of autonomous valet parking. The two-year project has been undertaken by a wide-ranging consortium and is being delivered in the Royal Borough of Greenwich, London as part of the UK Smart Mobility Living Lab.

Singapore autonomous vehicle testbed
Niels de Boer, programme director, CETRAN, Nanyang Technological University, SINGAPORE
Autonomous vehicles form part of Singapore's vision for urban mobility. Government agencies with support of industry and academia and led by the Land Transport Authority have established a testbed as well as a testing scheme to perform trials on public roads. The testing scheme consists of a qualification assessment, after which trials can commence and be continuously evaluated. This is done in a phased approach where test area and situational complexity increase over time. The goal is not only to test the technology itself, but also to develop the regulatory framework in parallel.

LUTZ pathfinder pods and urban laboratory in the UK
Dr Ujjar Bhandari, autonomous systems engineer (technologist), Transport Systems Catapult, UK
With the successful completion of the LUTZ Pathfinder project (led by Transport Systems Catapult), which accessed the feasibility of using fully automated vehicles to a last-mile solution in urban areas, the TSC is developing an urban laboratory facility that aims to provide open access to the LUTZ pathfinder pods/data to industry and academia for research and development purposes. This will provide a unique facility to trial autonomous systems in public shared spaces. This presentation will provide an update on the LUTZ pathfinder project, its outcomes and the opportunities and challenges facing the urban laboratory facility in Milton Keynes, UK.

What makes Vienna's test lab different: independent, dynamic anticipation, open innovation
Klaus Rosino, head of innovation, Smart Safety Solutions/Kuratorium für Verkehrssicherheit, AUSTRIA
The Vienna metropolitan region hosts an independent real-world living lab on sustainable automated driving. The global state of the art is tested in a real-world mixed-traffic testbed from the airport to the city centre conference arena. The focus is on dynamic interaction between pedestrians, two-wheelers and autonomous cars. The test track includes traffic lights as well as a significant element of the European C-ITS corridor from Rotterdam to Vienna. Momentum is added by the highly visible events accompanying Austria's presidency of EU-Council 2018 and Transport Research Arena 2018. This is being set up as a truly independent lab open to all by Smart Safety Solutions.

Sensor Verification and Validation
Room B

Static fusion of lidar and camera-based data for validation and reference purposes
Heiko Ruth, head of system development department, CMORE Automotive GmbH, GERMANY
Labelling processes and tools with mono ADAS camera systems are becoming increasingly accurate. But how accurate is the estimated distance in reality? A lidar reference system provides distance information in the centimetre range and is a huge support for the modern development cycle. Because the sensor under test (SUT) and the reference sensor have different coordinate systems, such a reference setup needs to have a coordinate transformation rule. This rule can be identified by a calibration method that allows the fusion of camera and lidar data in one picture.

Successful methods for open-road vehicle sensor validation
Simon Thompson, product manager, Oxford Technical Solutions, UK
Automotive R&D departments around the world need quick, reliable and repeatable methods to validate the accuracy of autonomous vehicle sensors, to <1cm accuracy. This presentation will share how 'ground truth' data can provide the additional positional reference that vehicle manufacturers need when mapping road feature positions and validating the accuracy of cameras and other sensors. We will share a number of current examples where this technology is being used effectively by OEMs in highly dynamic open-road conditions.

Synchronised emulation of ADAS sensor fusion targets – scenarios
Michael Konrad, CEO, Konrad GmbH, GERMANY
Today's ADAS end up in a combination of several different sensors, i.e. sensor fusion. Those systems become quite complex. To validate the functionality of those sensors, the manufacturer spends many kilometres on the street to get the right street scenarios. This presentation describes how sensor fusion systems (radar, lidar, camera and ultrasonic) can be validated with a sychronised target emulation for each single sensor, executed in the lab. The presentation will show the pro and cons of target emulation and will end with an overview of the sensor fusion emulation hardware and software.

A (really) new approach to lidar
Raul Bravo, CEO, Dibotics, FRANCE
Lidar sensors are becoming a key component of the autonomous car, which requires robust and affordable solutions to become a reality. The sensors are commonly used for several functions, including localisation, mapping, obstacle detection and classification. These are usually obtained when fusing the lidar data with IMU, encoders, radar and camera. What if a new approach to lidar allowed these features to be obtained directly and seamlessly from the lidar itself? What if this allowed for sensor fusion without needing calibration or synchronisation?

Road condition ahead - differentiating between the five road conditions
Dr Johan Casselgren, lecturer, Luleå University of Technology, SWEDEN
With more and more autonomous systems in our vehicles, road condition monitoring is becoming important. By using an optical road condition sensor in combination with a camera and an image-processing algorithm, a new method to classify road condition over the whole road lane has been developed. The new application can differentiate between the five road conditions – dry asphalt and asphalt covered by ice, snow, water and slush – in the whole lane at a distance of 1-15m in front of the vehicle, if the line of sight is clear and with the right illumination.

Sensor models for automated and assisted driving – a testing concept
Andreas Höfer, product manager, IPG Automotive GmbH, GERMANY
Automated driving functions require sensors to capture the vehicle’s environment. Therefore, suitable sensor models are indispensable for the simulation-aided development of these functions. This presentation introduces a comprehensive testing concept for sensors based on three classes of real-time capable sensor models: the first class comprises technology-independent sensor models for general function validation and fast prototype development. The second includes high-performance physical sensor models for robustness evaluation. The final class offers physically detailed sensor models including a raw data interface for component development and validation. The resulting flexible testing concept offers a suitable, resource-efficient sensor model for every use case.

Day 2

Wednesday 21 June

Real-World and Open-Road Testing
Room A

Moderator
Chris Reeves, Head of Connected & Autonmous Vehicles, Horiba Mira Ltd, UK

Multiple test providers in the Netherlands combine forces
Dirk-Jan de Bruijn, director, Traffic Innovatiecenter, NETHERLANDS
Offering a range of test facilities plays an increasingly dominant role in taking meaningful steps towards the implementation of new smart mobility applications with a significantly shorter time to market. Alongside living lab and simulation, this also covers real-life testing on the open road. With this in mind, we have taken the initiative to realise an integrated and cohesive offering with multiple test paths.

UK – an ecosystem for CAV testing and development
Michael Talbot, head of strategy, Centre for Connected & Autonomous Vehicles, UK
The UK has a well-documented history of invention and innovation. As the world steps into an era of large-scale disruption to the automotive and transport sectors, the UK intends to stay out in front by offering itself as the testbed for connected and autonomous vehicles. Building on recognised capability and bringing together expertise from all the key sectors, we are in the process of creating an innovation eco-system that is inclusive and drives the journey to an autonomous vehicle future.

Project SOHJOA – automated bus open-road piloting experiences
Harri Santamala, project director - Smarter Mobility, Metropolia UAS, FINLAND
Automated, SAE level 4, Easymile buses are being tested in open-road conditions in three Finnish cities. The presentation will focus on the necessary arrangements, best practices and lessons learned during the 2016 operation. It will discuss how the traffic adapted, and the typical issues the pre-mapped automated bus technology has to face in an open environment.

AD regulation review and new UTAC CERAM proving grounds
Alain Piperno, autonomous vehicle expert and project manager for testing and homologation, UTAC CERAM, FRANCE
As a worldwide reference in testing and homologation, a EuroNCAP-accredited lab and an active member of the new Industrial France Plan, UTAC CERAM will discuss: ADAS and AD regulation, standardisation and testing review; new UTAC CERAM ADAS and AD proving ground (end 2017); AD use cases, circuits and testing solutions.

Evaluation of the showcase with the Future Bus
Jeannet van Arum, director smart mobility, Provincie Noord-Holland, NETHERLANDS
In 2016 the Future Bus drove on the largest bus lane in Europe, between Haarlem and Schiphol. As road owner, the province of Noord-Holland facilitated the test. The evaluation will be presented.

Best Practices
Room A

Operational safety in autonomous vehicles
Peter Brink, principal engineer, PolySync Technologies Inc, USA
This presentation covers a proposed architecture for safety of an automated driving vehicle. Because of the current limitations in ISO 26262, this presentation includes a description of operational safety and how we use that in conjunction with functional safety to create an ADV that operates in a safe manner in a very wide variety of situations. This mechanism does not preclude using ISO 26262; it extends the development methodology beyond what is currently described and incorporates system-level analyses from the Verification and Validation Task Force under the On-Road Automated Driving (ORAD) Committee of the Society of Automotive Engineers (SAE).

Logical artificial intelligence in the control strategy of autonomous road vehicles
Dr Sergey Shadrin, scientific adviser, Mivar Ltd, RUSSIA
Mivar is a technological platform of logical artificial intelligence, based on expert systems. It is formalised as a network of objects (conditions, etc.) and links (rules) between them. Like rational human reasoning, it initially finds a way to solve the request and then applies a logical solution. One of the challenges in autonomous road vehicles is to make fast and correct decisions while driving, with further logging of each reasoned step. Traffic laws were described with Mivar techniques and implemented in the control system of an experimental autonomous car. Practical issues will be presented.

Who programs the programmer? A sceptic's view of validation processes
Nic Fasci, lead engineer - homologation, Tata Motors European Technical Centre, UK
Although the world is focused and committed to the deployment of automated vehicles, are the 'clever people' with the 1s and 0s the right people to be developing and evolving the algorithms for the vehicles? How can we be sure that the standard of driver training is adequate so that no gremlins or bad habits are installed from the outset during the test and development phase? Although the end goal is Level 5, humans are still key to the success of automated vehicles and if we get this wrong, the whole concept could be set back years.

Road safety for a network of connected vehicles
Eric Barbier, system safety engineer, Ricardo UK Ltd, UK
The connectivity considered for autonomous vehicles (V2V, V2I) brings a new challenge with regard to the road safety: how do we ensure that those vehicles will collectively behave in a safe manner, for the vehicle owner and also the other road users? How to approach and demonstrate the safety of connected vehicles? In this session the speaker will present an approach, from design to validation, of safety for a group of connected vehicles, effectively a 'system of systems'.

Scenario mining for virtual testing of automated driving
Sytze Kalisvaart, project manager, TNO Integrated Vehicle Safety, NETHERLANDS
The validation effort for automated driving functions is exploding. Based on fleet monitoring and machine learning techniques, TNO generates reusable road behaviour scenarios. The use of parametrisation enables characterisation of the scenarios and a significant data reduction. This provides a real-life driving behaviour model for control design, test case sampling and virtual testing. Various metrics are possible to monitor the coverage of the scenario base.

(The lack of) driver behaviour models in autonomous vehicles
Truls Vaa, senior research psychologist, Institute of Transport Economics (TOI), NORWAY
The average driver is extremely good at handling risks in traffic. The personal injury risk of a Norwegian driver is one accident per 425 years. That is what the systems of autonomous vehicles have to beat if they are aiming to reduce the number of accidents. The title of the presentation is an assertion that will be tested by evaluating publications that describe autonomous vehicles to an extent that enables appraisals of how the vehicle deals with the monitoring of risks and interactions with other road users, i.e. how driver behaviour is modelled in a given autonomous vehicle.

Test & Validation Methodology for Autonomous Vehicles
Room B

Terrabytes – the challenge of tomorrow's test tools
Joachim Fritzson, CEO, Zuragon, SWEDEN
To test collision avoidance systems in large quantities is a challenge. The large amounts of test data and log files are very rapidly reaching terrabytes in size, and the structures of the data are changing with the new ADAS and AD systems. To build up a test strategy, being able to handle the big data in an efficient and collaborative manner, using servers and cloud services will be essential. Zuragon presents a solution using industry computers or Android devices to enable the collection and post-analysis of today's and tomorrow's test scenarios.

Test and validation of V2X communication and applications
Axel Meinen, technical sales manager, S.E.A. Datentechnik GmbH, GERMANY
One part of the necessary technological basics for highly automated and autonomous driving is the intelligent wireless vehicle communication known as Vehicle-to-X (V2X) based on the IEEE 802.11p standard. Due to the safety-relevant functions of most of the components and applications, the mandatory and complex test and validation processes require an open and flexible platform, which allows developers to access, analyse and manipulate signals and data in the various implementation phases. This presentation introduces the available technologies and tools within the National Instruments platform for verification, test and validation for V2X of rf-parameters, protocols and applications.

Consistent and repeatable metrics to manage safety-critical software projects
Michal Jablonski, field application engineer, Vector Software, GERMANY
This presentation will discuss different metrics to manage development and testing activities to provide up-to-date information on the current release readiness. Selected information about code complexity, frequency of code changes, test case status and code coverage data and the right analysis, together with a graphical report, shows up possible bottlenecks. Resources can be spent in the right places and project leaders can make realistic assessments. Key insights into software quality will be given, plus how to use change-based tests, change impact analysis, test case maintenance and continuous testing to solve the recognised problems.

IIoT databus architecture for plug-and-play HIL testing platform
Bob Leigh, director of market development, Real-Time Innovations Inc, USA
Hear how one OEM is using a databus architecture to integrate multiple hardware and simulation components in a parallel and redundant HIL environment. Based on a single, standard-based framework, this worldwide test platform supports continuous testing through component failure using redundant and parallel simulations, and integrates multiple vendor products into one standard. The same databus architecture is used in autonomous and connected car networks and is a key part of any continuous testing process. Hear how a data-centric architecture, based on the DDS databus, can improve test performance and simplify development throughout the product lifecycle.

Resilience of automated systems: from base research to practice
Prof Frank Flemisch, branch head, Fraunhofer FKIE / RWTH Aachen, GERMANY
Vehicle automation is one way of using autonomous capabilities together with humans. While partially automated systems are already in serial production, highly and fully automated vehicles are under research and development. Resilience is the key ability to safely respond to non-normal events. This paper gives an introduction to the DFG System Ergonomics project, that deals with resilient design and testing of automated vehicle systems in normal operations, at system limits and failures. Some theory and the state of the research will be outlined, and examples given of how to practically apply this to the test and development of automated vehicles.

The building blocks of highly automated vehicles
Phil Magney, principal advisor, Vision Systems Intelligence LLC, USA
This session is an examination of the major hardware and software components used in the development of autonomous vehicle solutions. From sensors, to modules, to systems, VSI breaks down the major elements of an autonomous solution, including the latest trends in artificial intelligence (AI). Furthermore, we examine the challenges with sensors and sensor fusion and the need for localisation assets to improve motion planning. We will examine key hardware and software components (sensors, processors, modules and systems) as well as development tools, testing and validation requirements for functional safety purposes.

Lessons learned testing and validating unmanned vehicles
Jonathan Moore, chief engineer, Autonomous Solutions, USA
ASI has been automating ground vehicles including passenger cars, trucks, mining vehicles, military vehicles, security robots and indoor cleaning robots for over 16 years. In that time we have continuously improved our verification and validation capabilities to help speed up development and drive out defects and failures early. We've had some very interesting and difficult problems to solve, which our testing has been able to identify prior to product launch. This presentation shares some of those experiences and the real-life cases that can affect the behaviour of vehicles that don't have a driver.

High-speed multimedia datalink testing techniques
Mauro Arigossi, president & CEO, Alfamation SpA, ITALY
The rapid evolution and widening variety of automotive high-speed datalink solutions among sensors, central processing and displays requires sophisticated and flexible test approaches within the budget. A datalink technology overview is combined with testing principles and solutions for R&D and manufacturing.

Test specifications for highly automated driving function: highway pilot
Dr Hardi Hungar, senior researcher, German Aerospace Center (DLR), GERMANY
Based on approach and experiences in the ongoing PEGASUS project, a concept for deriving test specifications for highly automated driving functions is presented. A Level 3 system (highway pilot) is used to focus the research and demonstrate its results. It is proposed to formulate the test specification in terms of a catalogue of logical (abstract), parameterised scenarios. The pass/fail criterion relies on annotating the scenarios by criticality levels that must not be exceeded by the automation function. The paper also discusses ways of systematically deriving the test specification and how to perform tests.

Database of relevant traffic scenarios for highly automated vehicles
Andreas Pütz, senior engineer, FKA Forschungsgesellschaft Kraftfahrwesen mbH Aachen, GERMANY
The PEGASUS project is developing a validation framework for the sign-off process of highly automated vehicles. As one key element of this framework, a database of relevant traffic scenarios is implemented to reduce the necessary testing efforts. Incorporating input from different data sources (e.g. accident databases, field tests or driving simulator studies), the database compiles test specifications that can be used in the development and validation of automated driving functions. These test specifications aim to create high acceptance by their wide data basis and the collective development by the PEGASUS consortium.

Full-vehicle antenna and wireless OTA performance evaluation
Garth D'Abreu, director of automotive solutions, ETS-Lindgren, USA
The increasing use of antennas as the primary machine-to-environment interface is driving an increased reliance on the operation and performance of these devices. Autonomous features, especially those that have safety-related functionality, have to be rigorously evaluated at all stages of their operation. The over-the-air (OTA) portion is just one such stage, and the antenna-related performance of the communication channel here is often a critical part of the overall system sensitivity, reliability and operation. This presentation will highlight some measurement techniques, with the pros and cons of tests on modules and full-vehicle embedded systems.

Day 3

Thursday 22 June

Robust Cybersecurity for Autonomous and Connected Vehicles
Room A

Security unleashing the real power of connected cars
Martin Hunt, automotive industry practice lead, BT, GERMANY
Why do cars have brakes? Most would say for safety, to help us stop. But how about viewing them another way, as the things that enable you to drive faster? It’s the same with security. You can view it as a risk, something to protect yourself from, defend against. Or you can view it as an enabler. The real power of the connected car, and of the autonomous cars of the future, is the data they create. But these benefits will not be unleashed unless the security is in place to enable it to happen.

Effective security testing for automotive: 'CAN' we do it?
Dr Siraj Shaikh, Reader in Cyber Security, Coventry University, UK
We explore the challenges involved in security testing of the automotive controller area networks (CANs). Our interest lies in effective test case generation, threat assessment and compliance considerations, and AI techniques to characterise ECU-level behaviour for security anomalies.

Robust design and automated cybersecurity testing addressing cyber threats
Jürgen Ogrzewalla, vice president, FEV Europe GmbH, GERMANY
Cybersecurity presents design and validation challenges due to the increasing number of attack points in the vehicle and rapidly developing technology in the areas of connectivity, sophisticated infotainment and HMI, and autonomous driving. FEV provides cybersecurity services for these challenges through security management, security solutions and security testing. This presentation will go through some solution examples including hardware- and software-based approaches as well as an automated cybersecurity testing platform based on potential cyber threats.

Automotive threat modelling for security testing
Dr Paul Sanderson, senior technical specialist, SBD, UK
When conducting cybersecurity penetration testing, the time spent testing the system of interest is bounded. Security testing does not give 100% full coverage and unearth every single threat. Penetration testing, however, can be given a jump start by carrying out a threat model of the system under investigation to identify the potential vulnerabilities and attack points. This presentation introduces the methodology developed within the Secure Car Division of SBD, which has been successfully implemented on major OEM penetration testing programmes.

Security verification for the design of autonomous vehicles
Prof Ludovic Apvrille, assistant professor, Telecom Paristech, FRANCE
The rising wave of attacks on communicating embedded systems has exposed their users to risks of information theft, monetary damage and personal injury. The connectivity of autonomous vehicles will offer a tempting target to attackers. In this presentation, we summarise previous attacks on connected vehicles, and the additional security risks threatening autonomous vehicles. Furthermore, we propose countermeasures with secure architectures, and how to formally verify the design of such systems.

Security challenges on vehicular communication and standardisation issues
Dr Sang-Woo Lee, principal researcher, ETRI, KOREA
The presentation will offer an overview of secure vehicular communication technology including ITS and autonomous vehicle. It will also discuss standardisation issues in ITS security in SDOs such as IEEE and ITU-T, plus the current status of development on vehicular security technology.

Security concerns for current (and future) vehicle sensors
Harold Garza, research engineer, Southwest Research Institute, USA
The automotive landscape has drastically changed with respect to cybersecurity over the last few years, and autonomous vehicles will need to address the challenges presented by current and future autonomous vehicle technologies such as infotainment units, telematics and sensors. This presentation will begin by discussing the state of autonomous vehicle technologies that are already present in commercially available vehicles today. The presentation will then focus on the various sensor vulnerabilities that have been published so far, as well as some other vulnerabilities that could be exploited on the vehicle.

Using Simulation to Advance System Design and Validation
Room B

Autonomous vehicles: virtually validating functional, safety and performance requirements
Enguerrand Prioux, ADAS/AV product line manager, Siemens, FRANCE
The next challenge for numerical simulation, linked to autonomous vehicles, is unprecedented. The challenge is no longer to validate virtually a limited number of scenarios imposed by emissions standards but to cover hundreds of thousands scenarios. In terms of technology, it is a question of addressing physics (vehicle, sensors, traffic, road, etc.) and human modelling, to develop the artificial intelligence that will progressively replace the driver, to manage the chain from requirements to validation, automate and parallelise a large number of simulations, and trace and analyse their results. This vision will be illustrated through an adaptive cruise control example.

Test case synthesis and simulation for autonomous system validation
John Redford, VP architecture, Five AI Inc, UK
Sophisticated driving simulation is crucial to V&V for autonomous systems, enabling developers to parallelise the soak and regression testing of the system over extensive iterations of situation and hazardous events scenarios captured from the real world. Constrained random test generation tools can be integrated into these simulators to push that world into more entangled, complex states and generate hundreds of thousands of synthetic test cases for the run-time software. We present the key methods of synthesising new test cases and using them in simulation for autonomous system validation.

The use of 3D VR simulation in autonomous driving R&D
Dr Brendan Hafferty, general manager, Forum 8, UK
The presentation will discuss the application of real-time interactive 3D VR simulation and modelling technology in the research and development of autonomous vehicles, and the role of this software in road safety research. The paper will focus on the simulation of vehicle-to-vehicle and vehicle-to-infrastructure communication using the latest interactive 3D simulation software and a range of driving simulation hardware platforms. It will also highlight how the integration of data from third-party industry standard micro-simulation vehicle and pedestrian modelling software can add a further dimension to the overall simulation and design process.

Webots: a realistic automobile simulation software for autonomous vehicles
Dr Olivier Michel, CEO, Cyberbotics Ltd, SWITZERLAND
Cyberbotics Ltd has been developing the Webots software for more than 19 years. This 3D simulation software was originally designed to simulate mobile robots and therefore includes a series of sensors and actuators commonly found in robotics. Over the past three years, the software had been extended, in collaboration with PSA and Renault-Nissan, to support autonomous vehicle simulation, including physics simulation, automobile sensors (GPS, radar, lidar, cameras, ultrasonic, radio communications, etc.). Moreover, it includes tools to import maps from OpenStreetMap, to generate vehicle and pedestrian traffic (sumo), and to simulate changing weather conditions (rain, fog, night, dazzle).

Using driver-in-the-loop simulators for driverless vehicle developments
Phil Morse, technical liaison, Ansible Motion, UK
Recent advances in driver-in-the-loop (DIL) simulators allow experienced evaluation test drivers to interact with highly sophisticated physics models of vehicles and virtual environments. Such 'vehicle dynamics class' simulators, as tools, are aimed at providing convincing real-time feedback to drivers in order to sufficiently inform the task of driving a car. This presentation explores whether DIL simulators are useful for developing autonomous vehicles, when human drivers/occupants are not responsible for vehicle control tasks. Results from actual DIL experiments in the autonomous vehicle domain are discussed, as well as the procedures by which experimental validity can be defined.

Probabilistic sensor models for the simulation of ADAS sensor data
Dr Norman Mattern, CTO validation software, BaseLabs GmbH, GERMANY
The simulation of sensor data is one of the key challenges during development and validation of ADAS. Although ideal sensor models do not simulate any inaccuracies inherent in real sensor data, physical sensor models may not meet real-time requirements. The probabilistic sensor models of BaseLabs fill this gap by injecting typical random and systematic errors into the ideal data provided by simulation environments. The presentation provides an overview of the available probabilistic sensor simulation models and their technology. Furthermore, a workflow for the sensor model parameterisation based on the data of real sensors is presented.

High-performance radar target simulation for real-time HIL testing applications
Nicholas Keel, group manager - real-time test and monitoring, National Instruments, USA
As vehicle manufacturers develop the next generation of ADAS systems for semi-autonomous and autonomous vehicles, regulatory pressures and customer expectations will demand a higher level of safety than is required of today’s vehicles. Core technologies being developed include a blend of RF technologies, such as automotive radar, with embedded software in ADAS ECUs. HIL testing for ADAS and autonomous vehicles will need to include deterministic, low-latency RF radar target simulation to verify the proper operation of ADAS ECU software. Learn about the latest system architectures that will enable real-time embedded software testing for automotive radar sensors and ECUs.

Simulation-based forecasting of the impact of autonomous driving
Dr Jochen Lohmiller, manager PTV Vissim product management, PTV Group, GERMANY
The purpose of this paper is to discuss ways in which we can evaluate the impact of autonomous vehicles on the traffic flow by different penetration rates. In other words, how we can evaluate the coexistence of autonomous and conventional vehicles in the transition phase or the world after the transition phase with microscopic simulation tools. Results suggest that in high-speed environments (freeways) a huge increase in capacity can be achieved in dependency on headways and speed. In urban areas the performance of intersections will remain decisive.

Virtual testing to speed up self-driving automation
Rodolphe Tchalekian, pre-sales engineer, ESI GmbH, GERMANY
Self-driving automation deployment requires more and more physical test driving. Nevertheless, for addressing safety issues by improving the test coverage, simulation is becoming mandatory. Therefore, advanced sensors modelling, a key element in the simulation loop, has to reach higher levels of accuracy and reliability while including more complex physic aspects. In this paper we will introduce a new approach for object detection based on radar sensors by using new models taking into account the integration of the radar within the car and the radar signature of the target.

Integration of a simulation tool in ADAS MBT tool chain
Ahmed Yousif, software design engineer, Valeo Schalter und Sensoren GmbH, GERMANY
Sensor fusion is one of the major requirements of today’s complex advanced driving assistance systems (ADAS). To be able to master the challenge of the validation process of such complex features during product development, besides model-based testing (MBT), the use of a simulation tool becomes unavoidable, in particular when rapid prototyping is required. This paper presents the concept of integrating an MBT tool with a simulation tool to develop better software testing and validation processes by integrating real-life scenarios into the test process.

*This Programme may be subject to change.

Day 1

Tuesday 20 June

Keynote Presentations
Room A

Level 2 automation: lessons learned from crash investigation
Dr Ensar Becic, highway accident investigator / project manager, National Transportation Safety Board, USA
The first fatal crash involving an autonomous vehicle in the United States occurred in May 2016. The National Transportation Safety Board investigated this crash, in which a 2015 Tesla, operating in autopilot mode at that time, was travelling at 74mph on a state freeway, and struck and passed beneath the semitrailer of a combination vehicle that was completing a left turn. The autopilot did not detect the combination vehicle and the driver did not make any evasive manoeuvres. The investigation focused on (a) autopilot functionality and limitations, (b) driver trust and system misuse, (c) usability and standardisation of data.

Approaches for testing and type approval of automatic driving vehicles
Dr Houssem Abdellatif, global head autonomous driving & ADAS, TÜV SÜD Auto Service GmbH, GERMANY
Further road safety improvement is achievable through advanced driver assistance systems (ADAS) and automated driving (AD), which will result in autonomous driving in the future. However, common type approval procedures and testing approaches must be defined. The testing procedures based on physical testing seem to be insufficient to cover all the possible cases and thus to evaluate their efficiency and safety. Simulation methods seem to be the only feasible way to cover the huge number of possible scenarios. Physical tests will still be needed for verification and validation of these simulation models and setups.

Dynamic map layer technologies for intelligent automotive map data updates
Mike Tzamaloukas, vice president, autonomous drive and ADAS, and vice president, CoC navigation, Harman International Industries, Incorporated, GERMANY
As the number of sensors in cars increases, a technological solution for intelligent automotive map data updates can be used to create updates to existing road maps, and even support the necessary high-definition map content required for various safety and autonomy applications within the connected car. The presentation about dynamic map layer technologies will show how sensor fusion technology can be used to analyse safety and camera data to intelligently update map data.

Autonomous driving and dangerous weather-related road conditions?
Eric-Mark Huitema, global GM ground transport, IBM, The Weather Company, NETHERLANDS
An autonomously driving car cannot rely on the driver's interaction with the weather in its surroundings. However, weather is one of the main reasons for accidents and fatalities on safe roads. The solution must be a real-time and hyper-local weather interaction towards the car, so that the autonomous car can anticipate black ice, aquaplaning, thunderstorms, heavy rain or snow. This presentation will show how OEMs and aftermarket suppliers work with The Weather Company to make driverless interaction with the autonomous car possible.

Comprehensive modular verification and validation framework for automated systems
Andrea Leitner, project manager and research engineer, AVL List GmbH, AUSTRIA
This presentation aims to give an overview of the activities of ENABLE-S3, a large European project with around 71 partners from industry and research. The project’s goal is the provision of a comprehensive modular verification and validation framework to enable more efficient testing of automated systems. The framework does not only cover the testing platform, but also the testing methodology in order to reduce the number of tests. Based on requirements from 12 use cases in various domains, the result will be an approach to efficient and seamless scenario-based verification and validation in virtual, semi-virtual and real testing environments.

Open-Road and Real-World Testing
Room A

Moderator
Chris Reeves, Head of Connected & Autonmous Vehicles, Horiba Mira Ltd, UK

Navya worldwide development on private sites and open road
Christophe Sapet, CEO, Navya, FRANCE
The presentation will discuss autonomous vehicles at the heart of the smart city, and the benefits of autonomous vehicles on the first and last mile.

Development and testing of autonomous vehicles in urban environments
Dr Byoung-Kwang Kim, senior research engineer, Hyundai Motor Company, KOREA
We developed three autonomous driving IONIQs and two autonomous driving Tucson fuel cell electric vehicles. Users can call the vehicle like a taxi using smartphones and go to the destination autonomously. We tested them inside Hyundai’s Research and Development campus in Namyang, South Korea. Driving in the campus is a challenge because of the more than 10,000 employees on-site every day and almost urban environments. More than 1,000 volunteers, who are Hyundai's R&D engineers working in various divisions, experienced the full scenario of our autonomous vehicles. They assessed autonomous driving function and HMI.

GATEway – results from testing automated vehicles in Greenwich, London
Prof Nick Reed, academy director, TRL, UK
In this session, results from the GATEway project will be presented. This includes the testing of automated shuttle vehicles, a simulator study examining driver behaviour in the presence of automated vehicles, and the creation of a teleoperated vehicle capable of autonomous valet parking. The two-year project has been undertaken by a wide-ranging consortium and is being delivered in the Royal Borough of Greenwich, London as part of the UK Smart Mobility Living Lab.

Singapore autonomous vehicle testbed
Niels de Boer, programme director, CETRAN, Nanyang Technological University, SINGAPORE
Autonomous vehicles form part of Singapore's vision for urban mobility. Government agencies with support of industry and academia and led by the Land Transport Authority have established a testbed as well as a testing scheme to perform trials on public roads. The testing scheme consists of a qualification assessment, after which trials can commence and be continuously evaluated. This is done in a phased approach where test area and situational complexity increase over time. The goal is not only to test the technology itself, but also to develop the regulatory framework in parallel.

LUTZ pathfinder pods and urban laboratory in the UK
Dr Ujjar Bhandari, autonomous systems engineer (technologist), Transport Systems Catapult, UK
With the successful completion of the LUTZ Pathfinder project (led by Transport Systems Catapult), which accessed the feasibility of using fully automated vehicles to a last-mile solution in urban areas, the TSC is developing an urban laboratory facility that aims to provide open access to the LUTZ pathfinder pods/data to industry and academia for research and development purposes. This will provide a unique facility to trial autonomous systems in public shared spaces. This presentation will provide an update on the LUTZ pathfinder project, its outcomes and the opportunities and challenges facing the urban laboratory facility in Milton Keynes, UK.

What makes Vienna's test lab different: independent, dynamic anticipation, open innovation
Klaus Rosino, head of innovation, Smart Safety Solutions/Kuratorium für Verkehrssicherheit, AUSTRIA
The Vienna metropolitan region hosts an independent real-world living lab on sustainable automated driving. The global state of the art is tested in a real-world mixed-traffic testbed from the airport to the city centre conference arena. The focus is on dynamic interaction between pedestrians, two-wheelers and autonomous cars. The test track includes traffic lights as well as a significant element of the European C-ITS corridor from Rotterdam to Vienna. Momentum is added by the highly visible events accompanying Austria's presidency of EU-Council 2018 and Transport Research Arena 2018. This is being set up as a truly independent lab open to all by Smart Safety Solutions.

Sensor Verification and Validation
Room B

Static fusion of lidar and camera-based data for validation and reference purposes
Heiko Ruth, head of system development department, CMORE Automotive GmbH, GERMANY
Labelling processes and tools with mono ADAS camera systems are becoming increasingly accurate. But how accurate is the estimated distance in reality? A lidar reference system provides distance information in the centimetre range and is a huge support for the modern development cycle. Because the sensor under test (SUT) and the reference sensor have different coordinate systems, such a reference setup needs to have a coordinate transformation rule. This rule can be identified by a calibration method that allows the fusion of camera and lidar data in one picture.

Successful methods for open-road vehicle sensor validation
Simon Thompson, product manager, Oxford Technical Solutions, UK
Automotive R&D departments around the world need quick, reliable and repeatable methods to validate the accuracy of autonomous vehicle sensors, to <1cm accuracy. This presentation will share how 'ground truth' data can provide the additional positional reference that vehicle manufacturers need when mapping road feature positions and validating the accuracy of cameras and other sensors. We will share a number of current examples where this technology is being used effectively by OEMs in highly dynamic open-road conditions.

Synchronised emulation of ADAS sensor fusion targets – scenarios
Michael Konrad, CEO, Konrad GmbH, GERMANY
Today's ADAS end up in a combination of several different sensors, i.e. sensor fusion. Those systems become quite complex. To validate the functionality of those sensors, the manufacturer spends many kilometres on the street to get the right street scenarios. This presentation describes how sensor fusion systems (radar, lidar, camera and ultrasonic) can be validated with a sychronised target emulation for each single sensor, executed in the lab. The presentation will show the pro and cons of target emulation and will end with an overview of the sensor fusion emulation hardware and software.

A (really) new approach to lidar
Raul Bravo, CEO, Dibotics, FRANCE
Lidar sensors are becoming a key component of the autonomous car, which requires robust and affordable solutions to become a reality. The sensors are commonly used for several functions, including localisation, mapping, obstacle detection and classification. These are usually obtained when fusing the lidar data with IMU, encoders, radar and camera. What if a new approach to lidar allowed these features to be obtained directly and seamlessly from the lidar itself? What if this allowed for sensor fusion without needing calibration or synchronisation?

Road condition ahead - differentiating between the five road conditions
Dr Johan Casselgren, lecturer, Luleå University of Technology, SWEDEN
With more and more autonomous systems in our vehicles, road condition monitoring is becoming important. By using an optical road condition sensor in combination with a camera and an image-processing algorithm, a new method to classify road condition over the whole road lane has been developed. The new application can differentiate between the five road conditions – dry asphalt and asphalt covered by ice, snow, water and slush – in the whole lane at a distance of 1-15m in front of the vehicle, if the line of sight is clear and with the right illumination.

Sensor models for automated and assisted driving – a testing concept
Andreas Höfer, product manager, IPG Automotive GmbH, GERMANY
Automated driving functions require sensors to capture the vehicle’s environment. Therefore, suitable sensor models are indispensable for the simulation-aided development of these functions. This presentation introduces a comprehensive testing concept for sensors based on three classes of real-time capable sensor models: the first class comprises technology-independent sensor models for general function validation and fast prototype development. The second includes high-performance physical sensor models for robustness evaluation. The final class offers physically detailed sensor models including a raw data interface for component development and validation. The resulting flexible testing concept offers a suitable, resource-efficient sensor model for every use case.

*This Programme may be subject to change.

Day 2

Wednesday 21 June

Real-World and Open-Road Testing
Room A

Moderator
Chris Reeves, Head of Connected & Autonmous Vehicles, Horiba Mira Ltd, UK

Multiple test providers in the Netherlands combine forces
Dirk-Jan de Bruijn, director, Traffic Innovatiecenter, NETHERLANDS
Offering a range of test facilities plays an increasingly dominant role in taking meaningful steps towards the implementation of new smart mobility applications with a significantly shorter time to market. Alongside living lab and simulation, this also covers real-life testing on the open road. With this in mind, we have taken the initiative to realise an integrated and cohesive offering with multiple test paths.

UK – an ecosystem for CAV testing and development
Michael Talbot, head of strategy, Centre for Connected & Autonomous Vehicles, UK
The UK has a well-documented history of invention and innovation. As the world steps into an era of large-scale disruption to the automotive and transport sectors, the UK intends to stay out in front by offering itself as the testbed for connected and autonomous vehicles. Building on recognised capability and bringing together expertise from all the key sectors, we are in the process of creating an innovation eco-system that is inclusive and drives the journey to an autonomous vehicle future.

Project SOHJOA – automated bus open-road piloting experiences
Harri Santamala, project director - Smarter Mobility, Metropolia UAS, FINLAND
Automated, SAE level 4, Easymile buses are being tested in open-road conditions in three Finnish cities. The presentation will focus on the necessary arrangements, best practices and lessons learned during the 2016 operation. It will discuss how the traffic adapted, and the typical issues the pre-mapped automated bus technology has to face in an open environment.

AD regulation review and new UTAC CERAM proving grounds
Alain Piperno, autonomous vehicle expert and project manager for testing and homologation, UTAC CERAM, FRANCE
As a worldwide reference in testing and homologation, a EuroNCAP-accredited lab and an active member of the new Industrial France Plan, UTAC CERAM will discuss: ADAS and AD regulation, standardisation and testing review; new UTAC CERAM ADAS and AD proving ground (end 2017); AD use cases, circuits and testing solutions.

Evaluation of the showcase with the Future Bus
Jeannet van Arum, director smart mobility, Provincie Noord-Holland, NETHERLANDS
In 2016 the Future Bus drove on the largest bus lane in Europe, between Haarlem and Schiphol. As road owner, the province of Noord-Holland facilitated the test. The evaluation will be presented.

Best Practices
Room A

Operational safety in autonomous vehicles
Peter Brink, principal engineer, PolySync Technologies Inc, USA
This presentation covers a proposed architecture for safety of an automated driving vehicle. Because of the current limitations in ISO 26262, this presentation includes a description of operational safety and how we use that in conjunction with functional safety to create an ADV that operates in a safe manner in a very wide variety of situations. This mechanism does not preclude using ISO 26262; it extends the development methodology beyond what is currently described and incorporates system-level analyses from the Verification and Validation Task Force under the On-Road Automated Driving (ORAD) Committee of the Society of Automotive Engineers (SAE).

Logical artificial intelligence in the control strategy of autonomous road vehicles
Dr Sergey Shadrin, scientific adviser, Mivar Ltd, RUSSIA
Mivar is a technological platform of logical artificial intelligence, based on expert systems. It is formalised as a network of objects (conditions, etc.) and links (rules) between them. Like rational human reasoning, it initially finds a way to solve the request and then applies a logical solution. One of the challenges in autonomous road vehicles is to make fast and correct decisions while driving, with further logging of each reasoned step. Traffic laws were described with Mivar techniques and implemented in the control system of an experimental autonomous car. Practical issues will be presented.

Who programs the programmer? A sceptic's view of validation processes
Nic Fasci, lead engineer - homologation, Tata Motors European Technical Centre, UK
Although the world is focused and committed to the deployment of automated vehicles, are the 'clever people' with the 1s and 0s the right people to be developing and evolving the algorithms for the vehicles? How can we be sure that the standard of driver training is adequate so that no gremlins or bad habits are installed from the outset during the test and development phase? Although the end goal is Level 5, humans are still key to the success of automated vehicles and if we get this wrong, the whole concept could be set back years.

Road safety for a network of connected vehicles
Eric Barbier, system safety engineer, Ricardo UK Ltd, UK
The connectivity considered for autonomous vehicles (V2V, V2I) brings a new challenge with regard to the road safety: how do we ensure that those vehicles will collectively behave in a safe manner, for the vehicle owner and also the other road users? How to approach and demonstrate the safety of connected vehicles? In this session the speaker will present an approach, from design to validation, of safety for a group of connected vehicles, effectively a 'system of systems'.

Scenario mining for virtual testing of automated driving
Sytze Kalisvaart, project manager, TNO Integrated Vehicle Safety, NETHERLANDS
The validation effort for automated driving functions is exploding. Based on fleet monitoring and machine learning techniques, TNO generates reusable road behaviour scenarios. The use of parametrisation enables characterisation of the scenarios and a significant data reduction. This provides a real-life driving behaviour model for control design, test case sampling and virtual testing. Various metrics are possible to monitor the coverage of the scenario base.

(The lack of) driver behaviour models in autonomous vehicles
Truls Vaa, senior research psychologist, Institute of Transport Economics (TOI), NORWAY
The average driver is extremely good at handling risks in traffic. The personal injury risk of a Norwegian driver is one accident per 425 years. That is what the systems of autonomous vehicles have to beat if they are aiming to reduce the number of accidents. The title of the presentation is an assertion that will be tested by evaluating publications that describe autonomous vehicles to an extent that enables appraisals of how the vehicle deals with the monitoring of risks and interactions with other road users, i.e. how driver behaviour is modelled in a given autonomous vehicle.

Test & Validation Methodology for Autonomous Vehicles
Room B

Terrabytes – the challenge of tomorrow's test tools
Joachim Fritzson, CEO, Zuragon, SWEDEN
To test collision avoidance systems in large quantities is a challenge. The large amounts of test data and log files are very rapidly reaching terrabytes in size, and the structures of the data are changing with the new ADAS and AD systems. To build up a test strategy, being able to handle the big data in an efficient and collaborative manner, using servers and cloud services will be essential. Zuragon presents a solution using industry computers or Android devices to enable the collection and post-analysis of today's and tomorrow's test scenarios.

Test and validation of V2X communication and applications
Axel Meinen, technical sales manager, S.E.A. Datentechnik GmbH, GERMANY
One part of the necessary technological basics for highly automated and autonomous driving is the intelligent wireless vehicle communication known as Vehicle-to-X (V2X) based on the IEEE 802.11p standard. Due to the safety-relevant functions of most of the components and applications, the mandatory and complex test and validation processes require an open and flexible platform, which allows developers to access, analyse and manipulate signals and data in the various implementation phases. This presentation introduces the available technologies and tools within the National Instruments platform for verification, test and validation for V2X of rf-parameters, protocols and applications.

Consistent and repeatable metrics to manage safety-critical software projects
Michal Jablonski, field application engineer, Vector Software, GERMANY
This presentation will discuss different metrics to manage development and testing activities to provide up-to-date information on the current release readiness. Selected information about code complexity, frequency of code changes, test case status and code coverage data and the right analysis, together with a graphical report, shows up possible bottlenecks. Resources can be spent in the right places and project leaders can make realistic assessments. Key insights into software quality will be given, plus how to use change-based tests, change impact analysis, test case maintenance and continuous testing to solve the recognised problems.

IIoT databus architecture for plug-and-play HIL testing platform
Bob Leigh, director of market development, Real-Time Innovations Inc, USA
Hear how one OEM is using a databus architecture to integrate multiple hardware and simulation components in a parallel and redundant HIL environment. Based on a single, standard-based framework, this worldwide test platform supports continuous testing through component failure using redundant and parallel simulations, and integrates multiple vendor products into one standard. The same databus architecture is used in autonomous and connected car networks and is a key part of any continuous testing process. Hear how a data-centric architecture, based on the DDS databus, can improve test performance and simplify development throughout the product lifecycle.

Resilience of automated systems: from base research to practice
Prof Frank Flemisch, branch head, Fraunhofer FKIE / RWTH Aachen, GERMANY
Vehicle automation is one way of using autonomous capabilities together with humans. While partially automated systems are already in serial production, highly and fully automated vehicles are under research and development. Resilience is the key ability to safely respond to non-normal events. This paper gives an introduction to the DFG System Ergonomics project, that deals with resilient design and testing of automated vehicle systems in normal operations, at system limits and failures. Some theory and the state of the research will be outlined, and examples given of how to practically apply this to the test and development of automated vehicles.

The building blocks of highly automated vehicles
Phil Magney, principal advisor, Vision Systems Intelligence LLC, USA
This session is an examination of the major hardware and software components used in the development of autonomous vehicle solutions. From sensors, to modules, to systems, VSI breaks down the major elements of an autonomous solution, including the latest trends in artificial intelligence (AI). Furthermore, we examine the challenges with sensors and sensor fusion and the need for localisation assets to improve motion planning. We will examine key hardware and software components (sensors, processors, modules and systems) as well as development tools, testing and validation requirements for functional safety purposes.

Lessons learned testing and validating unmanned vehicles
Jonathan Moore, chief engineer, Autonomous Solutions, USA
ASI has been automating ground vehicles including passenger cars, trucks, mining vehicles, military vehicles, security robots and indoor cleaning robots for over 16 years. In that time we have continuously improved our verification and validation capabilities to help speed up development and drive out defects and failures early. We've had some very interesting and difficult problems to solve, which our testing has been able to identify prior to product launch. This presentation shares some of those experiences and the real-life cases that can affect the behaviour of vehicles that don't have a driver.

High-speed multimedia datalink testing techniques
Mauro Arigossi, president & CEO, Alfamation SpA, ITALY
The rapid evolution and widening variety of automotive high-speed datalink solutions among sensors, central processing and displays requires sophisticated and flexible test approaches within the budget. A datalink technology overview is combined with testing principles and solutions for R&D and manufacturing.

Test specifications for highly automated driving function: highway pilot
Dr Hardi Hungar, senior researcher, German Aerospace Center (DLR), GERMANY
Based on approach and experiences in the ongoing PEGASUS project, a concept for deriving test specifications for highly automated driving functions is presented. A Level 3 system (highway pilot) is used to focus the research and demonstrate its results. It is proposed to formulate the test specification in terms of a catalogue of logical (abstract), parameterised scenarios. The pass/fail criterion relies on annotating the scenarios by criticality levels that must not be exceeded by the automation function. The paper also discusses ways of systematically deriving the test specification and how to perform tests.

Database of relevant traffic scenarios for highly automated vehicles
Andreas Pütz, senior engineer, FKA Forschungsgesellschaft Kraftfahrwesen mbH Aachen, GERMANY
The PEGASUS project is developing a validation framework for the sign-off process of highly automated vehicles. As one key element of this framework, a database of relevant traffic scenarios is implemented to reduce the necessary testing efforts. Incorporating input from different data sources (e.g. accident databases, field tests or driving simulator studies), the database compiles test specifications that can be used in the development and validation of automated driving functions. These test specifications aim to create high acceptance by their wide data basis and the collective development by the PEGASUS consortium.

Full-vehicle antenna and wireless OTA performance evaluation
Garth D'Abreu, director of automotive solutions, ETS-Lindgren, USA
The increasing use of antennas as the primary machine-to-environment interface is driving an increased reliance on the operation and performance of these devices. Autonomous features, especially those that have safety-related functionality, have to be rigorously evaluated at all stages of their operation. The over-the-air (OTA) portion is just one such stage, and the antenna-related performance of the communication channel here is often a critical part of the overall system sensitivity, reliability and operation. This presentation will highlight some measurement techniques, with the pros and cons of tests on modules and full-vehicle embedded systems.

*This Programme may be subject to change.

Day 3

Thursday 22 June

Robust Cybersecurity for Autonomous and Connected Vehicles
Room A

Security unleashing the real power of connected cars
Martin Hunt, automotive industry practice lead, BT, GERMANY
Why do cars have brakes? Most would say for safety, to help us stop. But how about viewing them another way, as the things that enable you to drive faster? It’s the same with security. You can view it as a risk, something to protect yourself from, defend against. Or you can view it as an enabler. The real power of the connected car, and of the autonomous cars of the future, is the data they create. But these benefits will not be unleashed unless the security is in place to enable it to happen.

Effective security testing for automotive: 'CAN' we do it?
Dr Siraj Shaikh, Reader in Cyber Security, Coventry University, UK
We explore the challenges involved in security testing of the automotive controller area networks (CANs). Our interest lies in effective test case generation, threat assessment and compliance considerations, and AI techniques to characterise ECU-level behaviour for security anomalies.

Robust design and automated cybersecurity testing addressing cyber threats
Jürgen Ogrzewalla, vice president, FEV Europe GmbH, GERMANY
Cybersecurity presents design and validation challenges due to the increasing number of attack points in the vehicle and rapidly developing technology in the areas of connectivity, sophisticated infotainment and HMI, and autonomous driving. FEV provides cybersecurity services for these challenges through security management, security solutions and security testing. This presentation will go through some solution examples including hardware- and software-based approaches as well as an automated cybersecurity testing platform based on potential cyber threats.

Automotive threat modelling for security testing
Dr Paul Sanderson, senior technical specialist, SBD, UK
When conducting cybersecurity penetration testing, the time spent testing the system of interest is bounded. Security testing does not give 100% full coverage and unearth every single threat. Penetration testing, however, can be given a jump start by carrying out a threat model of the system under investigation to identify the potential vulnerabilities and attack points. This presentation introduces the methodology developed within the Secure Car Division of SBD, which has been successfully implemented on major OEM penetration testing programmes.

Security verification for the design of autonomous vehicles
Prof Ludovic Apvrille, assistant professor, Telecom Paristech, FRANCE
The rising wave of attacks on communicating embedded systems has exposed their users to risks of information theft, monetary damage and personal injury. The connectivity of autonomous vehicles will offer a tempting target to attackers. In this presentation, we summarise previous attacks on connected vehicles, and the additional security risks threatening autonomous vehicles. Furthermore, we propose countermeasures with secure architectures, and how to formally verify the design of such systems.

Security challenges on vehicular communication and standardisation issues
Dr Sang-Woo Lee, principal researcher, ETRI, KOREA
The presentation will offer an overview of secure vehicular communication technology including ITS and autonomous vehicle. It will also discuss standardisation issues in ITS security in SDOs such as IEEE and ITU-T, plus the current status of development on vehicular security technology.

Security concerns for current (and future) vehicle sensors
Harold Garza, research engineer, Southwest Research Institute, USA
The automotive landscape has drastically changed with respect to cybersecurity over the last few years, and autonomous vehicles will need to address the challenges presented by current and future autonomous vehicle technologies such as infotainment units, telematics and sensors. This presentation will begin by discussing the state of autonomous vehicle technologies that are already present in commercially available vehicles today. The presentation will then focus on the various sensor vulnerabilities that have been published so far, as well as some other vulnerabilities that could be exploited on the vehicle.

Using Simulation to Advance System Design and Validation
Room B

Autonomous vehicles: virtually validating functional, safety and performance requirements
Enguerrand Prioux, ADAS/AV product line manager, Siemens, FRANCE
The next challenge for numerical simulation, linked to autonomous vehicles, is unprecedented. The challenge is no longer to validate virtually a limited number of scenarios imposed by emissions standards but to cover hundreds of thousands scenarios. In terms of technology, it is a question of addressing physics (vehicle, sensors, traffic, road, etc.) and human modelling, to develop the artificial intelligence that will progressively replace the driver, to manage the chain from requirements to validation, automate and parallelise a large number of simulations, and trace and analyse their results. This vision will be illustrated through an adaptive cruise control example.

Test case synthesis and simulation for autonomous system validation
John Redford, VP architecture, Five AI Inc, UK
Sophisticated driving simulation is crucial to V&V for autonomous systems, enabling developers to parallelise the soak and regression testing of the system over extensive iterations of situation and hazardous events scenarios captured from the real world. Constrained random test generation tools can be integrated into these simulators to push that world into more entangled, complex states and generate hundreds of thousands of synthetic test cases for the run-time software. We present the key methods of synthesising new test cases and using them in simulation for autonomous system validation.

The use of 3D VR simulation in autonomous driving R&D
Dr Brendan Hafferty, general manager, Forum 8, UK
The presentation will discuss the application of real-time interactive 3D VR simulation and modelling technology in the research and development of autonomous vehicles, and the role of this software in road safety research. The paper will focus on the simulation of vehicle-to-vehicle and vehicle-to-infrastructure communication using the latest interactive 3D simulation software and a range of driving simulation hardware platforms. It will also highlight how the integration of data from third-party industry standard micro-simulation vehicle and pedestrian modelling software can add a further dimension to the overall simulation and design process.

Webots: a realistic automobile simulation software for autonomous vehicles
Dr Olivier Michel, CEO, Cyberbotics Ltd, SWITZERLAND
Cyberbotics Ltd has been developing the Webots software for more than 19 years. This 3D simulation software was originally designed to simulate mobile robots and therefore includes a series of sensors and actuators commonly found in robotics. Over the past three years, the software had been extended, in collaboration with PSA and Renault-Nissan, to support autonomous vehicle simulation, including physics simulation, automobile sensors (GPS, radar, lidar, cameras, ultrasonic, radio communications, etc.). Moreover, it includes tools to import maps from OpenStreetMap, to generate vehicle and pedestrian traffic (sumo), and to simulate changing weather conditions (rain, fog, night, dazzle).

Using driver-in-the-loop simulators for driverless vehicle developments
Phil Morse, technical liaison, Ansible Motion, UK
Recent advances in driver-in-the-loop (DIL) simulators allow experienced evaluation test drivers to interact with highly sophisticated physics models of vehicles and virtual environments. Such 'vehicle dynamics class' simulators, as tools, are aimed at providing convincing real-time feedback to drivers in order to sufficiently inform the task of driving a car. This presentation explores whether DIL simulators are useful for developing autonomous vehicles, when human drivers/occupants are not responsible for vehicle control tasks. Results from actual DIL experiments in the autonomous vehicle domain are discussed, as well as the procedures by which experimental validity can be defined.

Probabilistic sensor models for the simulation of ADAS sensor data
Dr Norman Mattern, CTO validation software, BaseLabs GmbH, GERMANY
The simulation of sensor data is one of the key challenges during development and validation of ADAS. Although ideal sensor models do not simulate any inaccuracies inherent in real sensor data, physical sensor models may not meet real-time requirements. The probabilistic sensor models of BaseLabs fill this gap by injecting typical random and systematic errors into the ideal data provided by simulation environments. The presentation provides an overview of the available probabilistic sensor simulation models and their technology. Furthermore, a workflow for the sensor model parameterisation based on the data of real sensors is presented.

High-performance radar target simulation for real-time HIL testing applications
Nicholas Keel, group manager - real-time test and monitoring, National Instruments, USA
As vehicle manufacturers develop the next generation of ADAS systems for semi-autonomous and autonomous vehicles, regulatory pressures and customer expectations will demand a higher level of safety than is required of today’s vehicles. Core technologies being developed include a blend of RF technologies, such as automotive radar, with embedded software in ADAS ECUs. HIL testing for ADAS and autonomous vehicles will need to include deterministic, low-latency RF radar target simulation to verify the proper operation of ADAS ECU software. Learn about the latest system architectures that will enable real-time embedded software testing for automotive radar sensors and ECUs.

Simulation-based forecasting of the impact of autonomous driving
Dr Jochen Lohmiller, manager PTV Vissim product management, PTV Group, GERMANY
The purpose of this paper is to discuss ways in which we can evaluate the impact of autonomous vehicles on the traffic flow by different penetration rates. In other words, how we can evaluate the coexistence of autonomous and conventional vehicles in the transition phase or the world after the transition phase with microscopic simulation tools. Results suggest that in high-speed environments (freeways) a huge increase in capacity can be achieved in dependency on headways and speed. In urban areas the performance of intersections will remain decisive.

Virtual testing to speed up self-driving automation
Rodolphe Tchalekian, pre-sales engineer, ESI GmbH, GERMANY
Self-driving automation deployment requires more and more physical test driving. Nevertheless, for addressing safety issues by improving the test coverage, simulation is becoming mandatory. Therefore, advanced sensors modelling, a key element in the simulation loop, has to reach higher levels of accuracy and reliability while including more complex physic aspects. In this paper we will introduce a new approach for object detection based on radar sensors by using new models taking into account the integration of the radar within the car and the radar signature of the target.

Integration of a simulation tool in ADAS MBT tool chain
Ahmed Yousif, software design engineer, Valeo Schalter und Sensoren GmbH, GERMANY
Sensor fusion is one of the major requirements of today’s complex advanced driving assistance systems (ADAS). To be able to master the challenge of the validation process of such complex features during product development, besides model-based testing (MBT), the use of a simulation tool becomes unavoidable, in particular when rapid prototyping is required. This paper presents the concept of integrating an MBT tool with a simulation tool to develop better software testing and validation processes by integrating real-life scenarios into the test process.

*This Programme may be subject to change.

 
Preliminary Conference Programme Online Now
Click Here to
view the Programme
10%
EARLY BIRD
DISCOUNT
BOOK NOW
Click here
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
 
Held Alongside
Autonomous Vehicle Technology World Expo 2017 Automotive Testing Expo 2017
Co-Located With:
 Autonomous Vehicle Design & Technology 2017 Autonomous Vehicle Software Symposium