Receding Horizon Defense Strategy for Reach-Avoid Games with Uncertainties via Pairwise Outcomes

This paper proposes a receding horizon defense strategy for a two-dimensional multiplayer reach-avoid game in which the defenders suffer from measurement uncertainties. In this game, the defenders aim at protecting a goal region from the attackers who strive to enter it. The whole game is decomposed to multiple subgames involving one defender and one attacker. For each subgame, a noisy attack region is proposed for the defender to characterize the set of positions that the attacker may reach before the defender. Then, the noisy interception point in the noisy attack region is introduced and used to design a defense winning strategy which prevents the noisy attack region from approaching the goal region. A maximum-matching defense strategy is designed for the defenders to capture as many attackers as possible. Simulation results are presented to validate our results.