Multiple-agent probabilistic pursuit-evasion games

Abstract

This paper addresses the control of a team of autonomous agents pursuing a smart evader in a non-accurately mapped terrain. By describing the problem as a partial information Markov game, we are able to integrate map-learning and pursuit. We propose receding horizon control policies, in which the pursuers and the evader try to respectively maximize and minimize the probability of capture at the next time instant. Because this probability is conditioned to distinct observations for each team, the resulting game is nonzero-sum. When the evader has access to the pursuers' information, we show that a Nash solution to the one-step nonzero-sum game always exists. Moreover, we propose a method to compute the Nash equilibrium policies by solving an equivalent zero-sum matrix game. A simulation example shows the feasibility of the proposed approach.

Improvement with Look-ahead on Cooperative Pursuit Games

“…A typical scenario is that a group of autonomous vehicles is tasked to suppress a cluster of adversarial moving targets, and it can be modeled naturally as a multi-player PE game. In [7]- [8], Hespanha et al formulated PE games under a probabilistic framework, in which greedy and one-step Nash equilibrium strategies are solved to maximize the probability of detecting evaders. Antoniades, Kim and Sastry studied several heuristic solutions in [9] and various implementation issues of PE strategies by a team of autonomous vehicles are addressed in [6] and [10].…”

Cooperative surveillance missions with multiple unmanned ground vehicles (UGVs)

“…Furthermore, a number of UAV surveillance papers, such as [7], [8] fall into this category. The pursuit-evasion problem, where a number of pursuers try to find an evader is sometimes also formulated in this way [9].…”

Optimal Agent Cooperation with Local Information

“…It suffices to selectively make observations on those parts of the environment relevant for the particular target acquisition problem at hand. This idea appears in [37,38,36] where the authors treat a pursuit-evader problem in an unknown environment. The pursuers observe locally, mapping relevant parts of the environment, while focusing on acquiring the evader.…”