Multi-robot Cooperative Pursuit Under Dynamic Environment

This paper mainly discusses the pursuit-evasion games, in which a team of autonomous mobile robots act as pursuers to pursue multiple moving targets cooperatively. The traditional ContractNet protocol is extended, including using case-based reasoning to reduce the scope of inviting bidding, and introducing assistant decision matrix to improve the alliance decision, so as to reduce the communication load during the task negotiation process. The concepts such as the alliance life value and penalty are also introduced. Based on these extensions, a kind of multi-robot cooperative pursuit algorithm that allows dynamic alliance is proposed. Simulation results show the feasibility and validity of the given algorithm.