Intercept Strategy for Maneuvering Target Based on Deep Reinforcement Learning

A novel interception strategy for high-speed maneuvering target based on reinforcement learning (RL) algorithm is proposed. With reshaped reward function and observation consisting solely of relative position information and path angle, the interception policy is trained by Deep Deterministic Policy Gradient (DDPG) algorithm. The results of training and test indicate that the proposed guidance law can approximate the optimal control model of the maneuvering target interception and show the superiority over traditional method.