Intelligent Control for Unmanned Flight Vehicles via Deep Reinforcement Learning

摘要

The problem of intelligent control for unmanned flight vehicles is studied in this paper. The linear model of unmanned flight vehicle is obtained by Jacobian linearization method according to the nonlinear model. The process of controller design can be divided into two steps. Firstly, to ensure the stability and prescribed performance of the closed loop system, the robust controller is given in terms of linear matrices inequalities based on robust control theory. Secondly, the intelligent controller is proposed to improve the transient performance via deep reinforcement learning. The parameters of controller can be tuned automatically in a neighborhood of robust controller's parameters. In the end, the simulation results are given to illustrated the effectiveness and superiority of the proposed method.