Decision-Making Method of Multi-UAV Cooperate Air Combat Under Uncertain Environment

Abstract

Multi-UAV cooperative air combat has attracted wide attention from relative scholars. However, the decision-making problem of UAV swarm confrontation under uncertain conditions makes it more difficult. In this article, a two-layer decision-making method, containing dynamic target assignment and distributed Monte Carlo tree search (MCTS), is proposed to address this issue. Additionally, the possibility degree function method of interval gray number is combined with a genetic algorithm to deal with uncertain information in an air combat environment. Specifically, considering the actual air combat scene, the target value factor is introduced in the target allocation process, and the dynamic target allocation mechanism is established to adjust the cluster combat strategy in real time. The experiments show that the proposed two-level decision-making method can effectively deal with the swarm air combat problem under uncertain environments. First, the improved genetic algorithm can solve the problem of target allocation in an uncertain environment and give the target allocation scheme in the current state. Moreover, the establishment of the dynamic target allocation mechanism makes the cooperative behavior of UAVs emerge in the swarm, which fully reflects the adversarial air combat.