Encoding Defensive Driving with Noncooperative Differential Games

David Fridovich-Keil, Chih-Yuan Chiu, Claire Tomlin

Abstract: This paper presents a novel formulation of robustness modeled on defensive driving. Like adversarial safety formulations, our method is also based in differential game theory; however, our approach is general-sum and applies to N-player scenarios. We present an overview of our method and preliminary results for two interactive driving examples in which the user may pre-specify the severity of avoidance maneuvers.

TL;DR: This paper introduces a novel formulation of safety inspired by defensive driving.

Keywords: Differential Game Theory, Motion Planning, Adversarial Robotic Safety