Scholarlink

<p>This paper addresses the trajectory tracking control problem of Unmanned Surface Vehicles (USVs) subjected to model uncertainties and external time-varying disturbances. To tackle these challenges, a novel model-free optimal control strategy is developed by integrating off-policy Integral Reinforcement Learning (IRL) with the Robust Integral of the Sign of the Error (RISE) technique. The off-policy IRL framework enables data-driven policy learning without prior knowledge of system dynamics, while the RISE-based component enhances robustness against both disturbances and uncertain dynamics. An exponential performance index is adopted to guide the optimal control design, and the stability of the overall control scheme is rigorously proven using Lyapunov theory. Simulation results demonstrate the superiority of the proposed method in terms of tracking accuracy and resilience when compared to traditional controllers. These findings highlight the potential of combining IRL and RISE in developing robust, model-free control solutions for complex marine systems operating in uncertain and dynamic environments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Model-Free Integral Reinforcement Learning Control Strategy with RISE for Unmanned Surface Vehicles Under Time-Varying Disturbances

  • 期刊论文

  • 发表日期：

    2026年3月2日

查看全文
  • Tran Thanh Tuan,
  • Vu Quoc Huy,
  • Bui Thanh Tan

摘要

This paper addresses the trajectory tracking control problem of Unmanned Surface Vehicles (USVs) subjected to model uncertainties and external time-varying disturbances. To tackle these challenges, a novel model-free optimal control strategy is developed by integrating off-policy Integral Reinforcement Learning (IRL) with the Robust Integral of the Sign of the Error (RISE) technique. The off-policy IRL framework enables data-driven policy learning without prior knowledge of system dynamics, while the RISE-based component enhances robustness against both disturbances and uncertain dynamics. An exponential performance index is adopted to guide the optimal control design, and the stability of the overall control scheme is rigorously proven using Lyapunov theory. Simulation results demonstrate the superiority of the proposed method in terms of tracking accuracy and resilience when compared to traditional controllers. These findings highlight the potential of combining IRL and RISE in developing robust, model-free control solutions for complex marine systems operating in uncertain and dynamic environments.