无人艇集群最优协同控制反演

To realize an optimal cooperative control strategy of unmanned surface vessel(USV) clusters under artificial control through data-driven learning, a linear quadratic closed-loop differential game inverse optimization algorithm is proposed. The algorithm can identify the cooperative strategy objective function according to the optimal system state and control input trajectories. In this study, an optimal feedback matrix is first identified based on the observed optimal system state and control input trajectories with additive white noise. The cooperative strategy objective function is then identified after solving the coupled algebraic Riccati equations derived from the necessary and sufficient conditions for Nash equilibria. The proposed inverse optimization algorithm can obtain the optimal cooperative strategy objective function to satisfy the given system state and control input trajectories. The objective functions identified by the inverse optimization algorithm can then be used to achieve an optimal cooperative control of USV clusters for specific task scenarios and provide new ideas and solutions for cluster adversarial games.