On October 8, 2020, Waymo opened its fully autonomous ride-hailing service to the general public in Phoenix. Right now, members of the public are hailing vehicles with no human driver controlling the car – either in the vehicle or remotely – to help them get to where they’re going as part of their everyday lives. This milestone is not just the output of many years’ experience in developing the Waymo Driver – it’s also the result of over a decade developing safety practices. Since its start in 2009 as the Google Self-Driving Car Project, Waymo has spent years refining a comprehensive set of robust methodologies that assess safety across its technology, operations and team culture, and ultimately guide the deployment and safe operations of the Waymo Driver. In this talk, Dr Trent Victor, Waymo's director of safety research and best practices will discuss the safety framework that guides Waymo's progress and is the first of its kind in the AV industry.

Volvo Cars has developed cars with ADAS since the beginning of 2000. The company's ambition is and always has been to make safe cars. How does the SOTIF ISO/PAS 21448 affect active safety development? Are we perhaps fulfilling the SOTIF PAS already? How did we handle non-ISO hazards before SOTIF PAS existed? The presentation will explore some use cases/examples where SOTIF affects the product.

To prove that automated driving systems (ADS) are safer than human drivers, it is suggested that they need to be driven for over 11 billion miles. A number of miles is not an appropriate metric and doesn’t guarantee absolute safety. This highlights the question: “How safe is safe enough?” To answer, we suggest a departure from the world of absolute safety to informed safety. A key aspect of informed safety includes an accurate and standardized definition of the operational design domain for an ADS, including conveying it to users, regulators and other stakeholders. The first step in all safety standards: ODD definition.

A new concept for an autonomous driving system is presented, which in particular protects other road users such as pedestrians and cyclists. Furthermore, the system automatically ensures compliance with traffic regulations in a smart city equipped with it. It goes far beyond what we have seen so far with lane-keeping or automatic braking systems. The concept of geofencing will be further developed with unprecedented granularity. Detailed maps of the traffic areas of a city with 4in resolution will be used.

The adoption and acceptance of autonomous vehicles in society depends heavily on the ability to have confidence in the systems and features providing autonomous functionality. This requires a significant focus on test, including test strategy, test workflow and test assets. However, testing autonomous systems is extremely challenging and novel, because these systems under test are developed and based on machine-learning algorithms and operate in a near-infinite state-space. This presentation will identify some strategies that vehicle manufacturers and suppliers employ to tackle this challenge, from simulation-heavy test to deployable hardware test solutions. It will cover the intricacies of simulation and areas of investment that have not been adequately addressed today.

To tackle the problem of verification and validation of autonomous vehicles, synthetic data generated in simulation environments can be employed to complement field testing, due to the fact that these environments are highly flexible and inexpensive. Although in recent years virtual environments have been employed by the community to develop and test ADAS/AD functionalities, such as AEB or ACC, there has been limited possibility to utilize synthetic data for sensing functions performance evaluation. This presentation will address the challenges and opportunities of simulation environments for verification and validation of camera-based sensing algorithms.

In an attempt to be as close to reality as possible, the most advanced AD and ADAS virtual test environments integrate real-world components. This HIL approach requires innovative new tools and methodologies to ensure smooth data flow between the virtual domain running synthetic scenarios, and the attached ECU real-world interfaces. This presentation will outline the challenges of quick and precise translations between physical and virtual domains within various HIL simulation setups for high-bandwidth video camera systems, and showcase how to overcome them with Xylon’s flexible and configurable single box solution.

This paper will highlight the challenges of going from hardware-in-the-loop (HIL) to vehicle road testing as there are gaps today that can only be resolved through live road testing. By enabling the inclusion of actual sensor modules and wireless interfaces, Autonomous Drive Emulation (ADE) will be a way to compress the time and cost of vehicle road testing. As an example, this paper will discuss some of the details of how this can be accomplished from a cellular vehicle-to-everything (C-V2X) perspective.

The challenge of autonomous vehicle validation relies on massive simulation, due to the vast number of kilometers to run in various road, traffic and weather conditions. Nevertheless, acceptance of the proposed automation system will have a heavy impact on the use of the system and the efficiency of handling it when sharing or taking back vehicle control. Driver-in-the-loop simulation will play an essential role, using high-performance driving simulators or dedicated configurations, including virtual reality and web-based online solutions. These simulation tools will be essential to complete efficient autonomous vehicle engineering design for driver acceptance and comfort.

Most new vehicles sold in 2018 (56%) had, as either a standard or optional feature, advanced driver assistance systems (ADAS) equipped with pedestrian automatic emergency braking (PAEB). However, current testing protocols implemented by safety organizations and programs, such as Euro NCAP, NCAP and IIHS largely do not evaluate the performance of PAEB in dark conditions without streetlights. This is true despite data from the National Highway Traffic Safety Association (NHTSA) showing that more pedestrian fatalities occur in dark conditions without streetlights than in daylight conditions. Additionally, independent tests conducted by NHTSA and the American Automobile Association (AAA) reveal that in every scenario considered, PAEB systems frequently fail to protect pedestrians in dark conditions. Within this context, to encourage improvement in PAEB performance and increase night-time pedestrian safety, this presentation includes the following sections: 1) A proposal for the expansion, in future testing protocols, of PAEB tests conducted in dark, night-time conditions. More precisely, we propose expanded testing in less than 1 lux (lx) ambient illuminance, using low-beam headlights, and without streetlights; 2) Results of Velodyne’s night-time PAEB tests comparing a lidar-based system with one example of current camera and radar-based technology; 3) An overview of Velodyne’s PAEB solution.

The technological challenges/opportunities of CAV/CAM and the roadmap to autonomy are maturing. However, the requirements to generate market pull and the skills required for the successful implementation and uptake of services are less well understood. For example, safe ongoing operation of vehicles needs new MOT tests addressing security, software and data privacy; new and disruptive ownership models present issues and opportunities regarding individual design and brand identity; trust and perception require more human-centered design for viable solutions. These and more questions need answering if economic, environmental and productivity projections are to be realized.

Adaptive driving beams incorporate technology that essentially removes the compromise between glare and illumination by eliminating the driver’s need to frequently switch between lower and upper beams. This technology automatically recognizes the lighting need and provides improved response time to pedestrians, animals and other unexpected objects in the road. Additionally, it is of great importance to the future of mobility by contributing to a more controlled optical environment for camera-based sensor technologies used for automated vehicles. But, as with all new technology, there are questions about how to adapt your testing program to accomplish the goals, as well as meet applicable standards. During this presentation, we will touch on a variety of topics, including adaptive driving beams and how they will affect the future of mobility, SAE J3069, alternative and upcoming regulations, and much more.

The Automotive Grade Linux (AGL) community consists of more than 150 companies across the automotive and tech industries who are working together to develop an open-source software platform for all in-vehicle applications from infotainment to autonomous driving. Sharing a single software platform across the industry decreases development times so OEMs and suppliers can focus on rapid innovation and bringing products to market faster. This talk provides an overview of AGL, production use cases including Toyota and Subaru, the project roadmap and how to get involved.

As the industry nears technical readiness in the development of automated vehicles, questions remain about how safe is safe enough. What does driving safely even mean? Many industry standards have been created that can assist in a safety assurance argument. But only one standard provides guidance on how we should understand what is reasonable to expect of others. This talk will explain the critical role that assumptions play in understanding what 'driving safely' means for an automated vehicle, and what work remains in order to fully realize a driverless future.

Automated driving solutions introduce new complexity into the development of embedded systems in cars. This complexity rises with each level of control and autonomy. The toolchain for such challenges is also complex and the integration of all the tools requires considerable effort without a real competitive advantage for the automated driving solution. Instead of solving these challenges alone, wasting lots of money along the way, Bosch's automated driving division has started an open-source community known as OpenADx. This talk will present the open-source approach, the current state of the community and the currently available solutions.