摘要
针对全垫升气垫船进坞过程出现的误差较大、速度较慢及易发生碰撞等问题,本文采用强化学习中的确定性策略梯度算法优化非线性自抗扰控制器设计的方法,并将优化后的自抗扰控制与PID控制相结合,通过对气垫船的艏向、航速与横向位移进行控制,实现了一种气垫船进坞的控制策略。通过仿真验证了此控制策略在对目标艏向快速跟踪的同时,提高了艏向控制对不确定性干扰的抵抗能力,实现了进坞过程的快速性与准确性。
To address the issues of large errors,slow response,and high-risk collision during the docking process of air cushion vehicles,a nonlinear active disturbance rejection controller was designed using the deterministic policy gradient algorithm in reinforcement learning.To realize a control strategy for the docking process that controls the heading,speed,and lateral displacement of vehicles,optimized active disturbance rejection and proportional integral derivative controls were combined.The simulation results demonstrate that this control strategy not only achieves rapid tracking of the target heading but also enhances the robustness of heading control against uncertain disturbances,thereby ensuring both the speed and precision of the docking process.
作者
王元慧
张峻恺
吴鹏
WANG Yuanhui;ZHANG Junkai;WU Peng(College of Intelligent Science and Engineering,Harbin Engineering University,Harbin 150001,China)
出处
《哈尔滨工程大学学报》
北大核心
2025年第7期1340-1348,共9页
Journal of Harbin Engineering University
基金
国家自然科学基金项目(52471377)。
关键词
全垫升气垫船
非线性自抗扰控制
强化学习
确定性策略梯度
神经网络
PID控制
艏向控制
航速控制
外界扰动
air cushion vehicle
nonlinear active disturbance rejection control
reinforcement learning
deterministic strategy gradient
neural network
PID control
heading control
speed control
external disturbance
