Method for optimizing cooperative cruising strategy of surface unmanned multi-agent system based on digital twin

To address the issue of insufficient autonomy and coordination in cooperative cruising of surface unmanned multi-agent systems, this study proposes an optimization method for cooperative cruising strategies based on a digital twin framework.

First, a digital twin model of the physical unmanned multi-agent system is constructed, and a mathematical model of cooperative cruising is established to analyze the motion characteristics of the cooperative cruising process. Then, considering the mutual influences and cooperative relationships among the agents, a proximal strategy optimization algorithm is employed to enhance the cooperative cruising efficiency of the unmanned multi-agent system. Finally, the proposed method is verified using the digital twin model for surface multi-agent systems, demonstrating the improvement in autonomous multi-agent training performance.

Compared with the multi-agent deep deterministic policy gradient (MADDPG) algorithm, the proposed muti-agent proximal policy optimization (MAPPO) algorithm achieves a 14.7% improvement in average reward and demonstrates more stable convergence. The unmanned agents are capable of forming a uniformly distributed formation centered around the patrol target, thereby providing more comprehensive information for cooperative cruising.

The study provides significant theoretical and practical support for optimizing cooperative cruising strategies of unmanned multi-agent systems operating on the water surface.