Cross-cycle iterative unmanned aerial vehicle reentry guidance based on reinforcement learning

The traditional predictive correction algorithm requires a large number of iterative calculations for the predicted trajectory, which greatly occupies a large amount of computing resources, so that the real-time solution of the guidance command can not be guaranteed, and the guidance accuracy will have a large impact. And the prediction correction guidance requires the algorithm to have the ability of selfadaptation and intelligent learning. Therefore, this paper proposes a cross-cycle iterative hypersonic UAV predictive correction guidance method based on reinforcement learning. The parametric control variable (CVP) method is used to construct the parametric model of the guidance command. The actor-critic-based reinforcement learning method is used to solve the guidance command in real time, and the guidance information is effectively transmitted in the adjacent guidance solution cycle. The guidance error converges to within the allowable accuracy range during the cross-cycle iteration. Monte Carlo simulation shows that the proposed method has good adaptability to initial conditions and flight parameter uncertainty, and can guarantee the real-time performance of the guidance command while achieving high-precision guidance.