Cooperative Air Defense Operations Technology Based on Reinforcement Learning

In response to the urgent demand for intercepting low-altitude unmanned aerial vehicles (UAVs) in vicineland security system, a game framework is proposed for the coordinated air defense operations of anti-aircraft (AA) artillery against unmanned air vehicles in low altitude, and the simulation model of the offensive and defensive sides is built. As for the interception problems of low altitude UAV in vicinagearth security scenario, a game confrontation evolution strategy framework of "artillery-plane" is proposed for the first time based on multi-agent reinforcement learning. By designing guiding reward functions, the difficulty of reward sparsity is addressed. The simulation results demonstrate the high effectiveness of UADA in intercepting UAVs with evolving attack strategies. This provides robust technical support for the intelligent development of low-altitude air defense in future vicineland security scenarios.