PERFORMANCE OF A SWITCHING CONTROLLER FOR PURSUIT-EVASION SCENARIOS WITH NOISY MEASUREMENTS

摘要

This work presents an application of a switched controlrnlaw for the pursuing agent in the Homicidal Chauffeurrngame. Past work by the authors has focused on the developmentrnof a fast decision making algorithm for pursuitevasionrnscenarios based on partitioning the state space withrnregional objective functions chosen from a library of availablernstrategies. The algorithm determines where, in termsrnof the state space, that each objective should be applied. Inrndoing so, what is ordinarily a minimax problem becomes arnsimpler optimal control problem. One advantage of thisrnstrategy is that it is possible to rapidly update a controlrnlaw based on measurements of another agent’s dynamicrncapabilities acquired during play. In this work, we detailrnthe construction of a switched controller for the HomicidalrnChauffeur game. The pursuer initially has no knowledge ofrnthe evader’s dynamics and must use speed measurements tornupdate its control law during the game. Results outlined inrnthis paper show a 25.6% increase in capture time comparedrnto the case with perfect information with no decrease in thernnumber of states that ultimately lead to capture.