Performance-prescribed reinforcement learning trajectory-tracking control of an unmanned surface vehicle

Objectives This paper addresses the oscillatory instability in trajectory tracking errors of unmanned surface vehicle (USV) caused by propulsion saturation under complex navigation conditions, by proposing a prescribed performance-based reinforcement learning optimal control method. Methods First, a novel saturation function is introduced to handle USV input saturation. Second, an improved prescribed performance control scheme is designed, which constrains tracking error convergence via an asymmetric performance boundary, relaxing stringent dependence on initial error states. Then, a reinforcement learning optimization mechanism based on an Actor‑Critic framework is constructed to iteratively learn the optimal control policy and value function, achieving performance optimization under state constraints. Finally, the stability of the closed‑loop tracking system is rigorously proven via Lyapunov theory. Results Numerical simulations performed on the KVLCC2 tanker model demonstrate that the proposed method effectively handles the trajectory tracking problem subject to saturation constraints, with all tracking errors strictly confined within the prescribed performance boundaries. Conclusions The study provides a new solution for high‑performance tracking control of constrained USVs and exhibits practical value for engineering applications.