Learning-Based Cooperative Control Method of High-Speed Aircraft Swarm

As an emerging technology, the collaborative control of aircraft swarms can effectively improve the capabilities of aircraft swarms in target search, coordinated strikes and other application scenarios. Aiming at the problem of collaborative control of high-speed aircraft swarms, this paper proposes a high-speed aircraft swarm control method based on leader-follower architecture and deep reinforcement learning (DRL) method. The high-speed aircraft swarm control problem is decomposed into the threat avoidance control of the leader and the state consistency control of the followers, and the control strategy of the followers is constructed based on the DRL method. In this way, the complex swarm cooperative control problem is transformed into a simple leader-follower control problem, and a simple and effective high-speed aircraft swarm control strategy can be obtained. Simulation results show that the proposed flocking control strategy can control high-speed aircraft swarms to form a tight and consistent flock in a complex flight environment.