Prescribed Performance Sliding Mode Control with Parameter Tuning Based on DDPG for RLV in Reentry Phase

In this paper, the attitude control algorithm of reusable launch vehicle (RLV) reentry phase is studied. Firstly, to improve the transient performance while maintaining robustness, a control algorithm combining global fast terminal sliding mode and prescribed performance control (PPC) is designed for the RLV attitude control in reentry phase. In the proposed algorithm, a novel fast reaching law is utilized to improve the convergence speed of the sliding mode method. Furthermore, aiming at the difficulty of parameter tuning in the proposed algorithm, a parameter optimization strategy based on deep deterministic policy gradient (DDPG) is designed to tune the controller parameters online. Finally, the effectiveness of the proposed algorithm is verified by simulation results.