Nash-Minmax Capturing Strategy for Multiagent Pursuit-Evasion Differential Games Under Directed Graphs

In this article, a heterogeneous multiagent pursuit-evasion (MPE) differential gaming problem is addressed. Therein, the pursuers aim to cooperatively capture the sole escaping evader. The theoretical analytical challenge lies in designing the Nash-minmax gaming strategy for such MPE games under directed graphs with external disturbances. Intra-alliance cooperation (pursuers and their peer neighbors) is regarded as a nonzero-sum game, while interalliance confrontation (multiple pursuers versus one evader) is treated as another zero-sum game. To tackle this issue, a hierarchical gaming framework is established to solve the associated coupled Hamilton-Jacobi-Isaacs (HJI) equation and accordingly derive the associated distributed gaming protocol. Finally, both numerical simulations and lake-based multi-autonomous surface vehicle (ASV) gaming experiments are conducted to verify the effectiveness of the present MPE gaming strategy.