Design and Verification of UAV Maneuver Decision Simulation System Based on Deep Q-learning Network

Abstract

With the development of artificial intelligence technology and the demand of air combat, autonomous maneuver decision of UAV has become a popular research direction. Scholars at home and abroad have researched on autonomous air combat maneuver decision of UAV in depth based on various technologies and have achieved some results, among which the maneuver decisions based on the reinforcement learning are more efficient. But at present, the simulation targets used to test the effectiveness of the UAV autonomous maneuver decision method are relatively fixed, which cannot reflect the complexity of the enemy's maneuver strategy under real air combat conditions. This paper designs and develops a set of UAV autonomous maneuver decision man-machine air combat system based on deep Q-learning network, which is built from three subsystem: air combat environment simulation subsystem, the manned aircraft operation simulation subsystem and UAV self-learning subsystem. It has been verified that the man-machine air combat system proposed in this paper can effectively simulate real air combat and verify the effectiveness of the UAV autonomous maneuver decision.