A resilient active target defense strategy using multiple UAVs with nonlinear dynamics

Abstract

In this study, a resilient defense strategy (RDS) is proposed for a team of unmanned aerial vehicles (UAVs) with nonlinear dynamics to protect the target in a multiple defender multiple attacker active target defense differential game. To keep a defender’s dominance over an attacker, a control barrier function (CBF)-based controller is designed based on the geometric relationships between players’ dominant regions formed by the defender-attacker game and the attacker-target game, which is applicable to players with nonlinear control-affine dynamics. This controller integrates a forward-invariant constraint that connects nonlinear players’ control variables to their dominant regions, ensuring the defense winning condition in a 1-defender-1-attacker game. To dynamically allocate the defenders against potential intelligent attackers with evasive maneuvers in a multiple-defender-multiple-attacker game, resilience is integrated with a designed reallocation mechanism to form the RDS. When any of the allocated defenders no longer takes the dominant position, a reallocation of the defenders is triggered. In terms of the defense winning rate, the results show that the RDS outperforms the distance-based task allocation method with a pure pursuit controller. Moreover, a higher defender-to-attacker speed ratio helps improve the winning rate of the defenders, while the disadvantaged maneuverability (angular turn rates) of the defenders does not degrade the winning rate when there are sufficient defenders.