Path Planning and Control of Autonomous Vehicles in Dynamic Reach-Avoid Scenarios

Abstract

This paper addresses the dynamic adversarial reach-avoid problem in which the attacker aims at visiting a fixed target while avoiding the other vehicle (defender), which is trying to capture the attacker. We address the problem of control and path planning for both the attacker and the defender vehicles, capture the assumptions and requirements of this dynamic complex problem in the form of an LTL formula, and construct a hybrid controller which generates wining strategies for the vehicle under control while respecting its physical constraints.