A Proposed Safety Case Framework for Automated Vehicle Safety Evaluation
Jeffrey Wishart, Junfeng Zhao, Braeden Woodard, Gavin O'Malley, Hencong Guo, Shujauddin Rahimi, Sunder Swaminathan
Abstract
Automated driving system (ADS)-equipped vehicles (AVs) are currently being developed and deployed on public roads. While the benefits to public safety (among other potential benefits) are promising, the risk that AVs pose to public safety is not yet understood. The Automated Vehicle Test and Evaluation Process (AV - TEP) mission initiated by Science Foundation Arizona in collaboration with Arizona State University is intended to provide a framework for an AV developer or third-party evaluator (such as a regulator) to use in order to provide evidence that the AV is safe for its intended implementation. The AV-TEP framework uses the safety case concept that is widely accepted in the AV industry and others, and consists of three pillars: (1) Safety Management System (SMS), (2) Design Methods, and (3) Scenario-Based Testing. The three pillars are described in detail, and the validation methodology used is outlined. Finally, the various paths to industry acceptance of the AV-TEP framework are described, which builds upon existing work and research wherever appropriate.