Dense reinforcement learning for safety validation of autonomous vehicles.

One critical bottleneck that impedes the development and deployment of autonomous vehicles is the prohibitively high economic and time costs required to validate their safety in a naturalistic driving environment, owing to the rarity of safety-critical events1. Here we report the development of an intelligent testing environment, where artificial-intelligence-based background agents are trained to validate the safety performances of autonomous vehicles in an accelerated mode, without loss of unbiasedness. From naturalistic driving data, the background agents learn what adversarial manoeuvre to execute through a dense deep-reinforcement-learning (D2RL) approach, in which Markov decision processes are edited by removing non-safety-critical states and reconnecting critical ones so that the information in the training data is densified. D2RL enables neural networks to learn from densified information with safety-critical events and achieves tasks that are intractable for traditional deep-reinforcement-learning approaches. We demonstrate the effectiveness of our approach by testing a highly automated vehicle in both highway and urban test tracks with an augmented-reality environment, combining simulated background vehicles with physical road infrastructure and a real autonomous test vehicle. Our results show that the D2RL-trained agents can accelerate the evaluation process by multiple orders of magnitude (103 to 105 times faster). In addition, D2RL will enable accelerated testing and training with other safety-critical autonomous systems.

Systematic review: Automated vehicles and services for people with disabilities.

People with disabilities face many travel barriers. Autonomous vehicles and services may be one solution. The purpose of this project was to conduct a systematic review of the grey and scientific literature on autonomous vehicles for people with disabilities. Scientific evidence (n = 35) was limited to four observational studies with a very low level of evidence, qualitative studies, reviews, design and model reports, and policy proposals. Literature on older adults was most prevalent. Grey literature (n = 37) spanned a variety of media and sources and focuses on a variety of disability and impairment types. Results highlight opportunities and barriers to accessible and usable AVs and services, outline research gaps to set a future research agenda, and identify implications for policy and knowledge translation. People with disabilities are a diverse group, and accessible and usable design solutions will therefore need to be tailored to each group's needs, circumstances, and preferences. Future research in diverse disability groups should include more participatory action design and engineering studies and higher quality, prospective experimental studies to evaluate outcomes of accessible and usable AV technology. Studies will need to address not only all vehicle features but also the entire travel journey.