Research on Control Method of Aerial Refueling Aircraft Based on Reinforcement Learning

In the design of the precision control system of traditional aerial refueling aircraft, there are often coupling characteristics, which make its aerodynamic characteristics very complex and greatly affect the stability and reliability of the aircraft control system. In order to bypass the tedious traditional decoupling process, this paper uses the deterministic strategy search algorithm (DDPG) to train the neural network into an intelligent controller, and carries out simulation tests in the aircraft aerial refueling stage and in the environment with random wind field interference. The results verify the feasibility of intelligent control of aerial refueling aircraft.