RL-based Prescribed Performance Path Following Control for USVs with Intervened Triggering Guidance
Zhihao Li, Guoqing Zhang, Jiqiang Li, Xuanzhi Zhu · IFAC-PapersOnLine · 2025
This paper proposes a novel guidance and control scheme for unmanned surface vessels (USVs) that perform path following tasks in marine environments subject to external disturbances. By integrating a triggering boundary rule and a saturation compensation function, an intervened triggering guidance method is developed to mitigate the issue of communication redundancy and input saturation caused by irrational heading signals. To enhance the control accuracy and system responsiveness, a prescribed performance control (PPC) algorithm is designed to ensure both transient and steady-state control performance. Furthermore, reinforcement learning (RL) in the form of actor-critic neural networks (AC-NNs) is incorporated to address the challenges posed by model uncertainties and external disturbances. Ultimately, the stability analysis and numerical experiments are conducted to validate the feasibility and robustness of the proposed scheme.