Work

Event-Triggered Optimal Path-Following Control for Wind-Assisted Autonomous Surface Vehicles via Actor–Critic Reinforcement Learning

Year: 2025

Type: article

Abstract: This paper proposes an enhanced event-triggered optimal control scheme integrated with reinforcement learning (RL) for wind-assisted autonomous surface vehicles (WAASVs), aiming to ensure the safety and energy efficiency of marine path-following missions. To address the uncertainties arising from the dynamic model and time-varying external disturbances, a reinforcement learning approach based on the architecture of actor–critic neural networks (AC-NNs) is employed to generate control signals without relying on precise model knowledge while minimizing path deviation and energy consumption. Furthermore, an integral event-triggered control (IETC) algorithm is developed to dynamically adjust the control signal updates according to the system output errors, which offers a promising solution to prevent excessive mechanical wear of the actuators. The stability of all error variables is rigorously analyzed using the Lyapunov theory. Finally, two simulation experiments on the rotor-assisted vehicle are performed to validate the superior tracking performance and practical applicability of the proposed algorithm.

Source: Journal of Marine Science and Engineering

DOI: 10.3390/jmse13112117

Authors: Zhihao Li, Guoqing Zhang, Peng Liu

Institution: Dalian Maritime University

Language: English

Topic: Adaptive Dynamic Programming Control

Subfield: Computational Theory and Mathematics

Field: Computer Science

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $3,137

Funders: National Natural Science Foundation of China, Fundamental Research Funds for the Central Universities