High-Fidelity Decision-Making and Simulation for Cooperative Autonomous Air Combat Considering the Effect of Flight Controller

Abstract

In the past decade, how to improve the fidelity of maneuver gaming and the synergy of target allocation has become a key issue in cooperative autonomous air combat researches. To address the problem, this study proposes a maneuver decision-making algorithm based on an optimized dynamic Bayesian network and a target allocation decision-making algorithm based on an optimized hybrid particle swarm optimization. In maneuvering decision-making, the state transition’s reliability and the air combat’s autonomy are enhanced through considering the effect of sliding mode control. The Bayesian network is improved through introducing a strategy for dynamic prior probability updating. The computation is reduced and the efficiency is increased through pruning the minimax search tree according to visual prediction. In target allocation decision-making, the algorithm’s convergence speed is greatly accelerated and the solution’s global optimality is improved through introducing immigrant particles. The algorithm’s application scope is expanded through proposing a solution principle about unequal quantity combat situations. Furthermore, the end criterion is specially designed to fit real-world combats through introducing a fire control. The simulation results show that the designed decision-making algorithms are more effective in solving the problem of cooperative autonomous air combat, which indicates that the various improvements introduced in this study are reasonable and effective.

Keywords

• Cooperative autonomous air combat
• maneuver decision-making
• target~allocation decision-making
• sliding mode controller
• dynamic Bayesian network
• hybrid particle swarm optimization