A Modeled Approach for Online Adversarial Test of Operational Vehicle Safety

A scenario-based test of operational vehicle safety presents a set of principal other vehicle (POV) trajectories that seek to enforce the subject vehicle (SV) into a certain safety-critical situation. Current scenarios are mostly (i) statistics-driven: inspired by human driver crash data, (ii) deterministic: POV trajectories are pre-determined and are independent of SV responses, and (iii) overly simplified: defined over a finite set of actions performed at the abstracted motion planning level. Problems thus arise, making the scenario-based test (i) lack of severity guarantees, (ii) easy to hack with intelligent driving policies, and (iii) inefficient in producing safety-critical instances with limited and expensive testing effort. In this paper, the authors propose a model-driven feedback control policy for multiple POV's to propagate efficient adversarial trajectories online while respecting traffic rules and other concerns formulated as admissible state-action space. The proposed approach is formulated in an anchor-template hierarchy structure, with the template model planning inducing a theoretical SV capturing guarantee under standard assumptions. The planned adversarial trajectory is then tracked by a lower-level controller applied to the full-system or anchor model. The effectiveness of the proposed methodology is illustrated through various numeric examples with SV controlled by parameterized self-driving policies and human drivers.