基于实时反馈强化学习神经网络的船舶艏摇智能控制研究
Research on Intelligent Control of Ship Yaw Based on Real-time Feedback Reinforcement Learning Neural Network
宋伟伟 1徐跃宾 2段学静 1巩方超 1崔英明2
作者信息
- 1. 山东省船舶控制工程和智能系统工程技术研究中心,山东威海 264300;威海海洋职业学院,山东威海 264300
- 2. 威海海洋职业学院,山东威海 264300
- 折叠
摘要
文章提出了一种基于实时反馈强化学习神经网络控制的船舶艏摇智能控制方法.该方法将神经网络的非线性建模和强化学习的自适应控制技术相结合,能够实现对船舶航行过程中舵角的精确控制.并将PID控制算法、模型预测控制算法和实时反馈强化学习神经网络控制算法进行对比分析,仿真实验结果表明,后者在控制效果和稳定性方面均优于前两种方法,能够有效地提高船舶航行过程中舵角的控制精度和鲁棒性.
Abstract
This paper proposes an intelligent control method for ship yaw based on real-time feedback reinforcement learning neural network control.This method combines nonlinear modeling of neural networks with adaptive control technology of reinforcement learning to achieve precise control of rudder angle during ship navigation.And the PID control algorithm,model prediction control algorithm,and real-time feedback reinforcement learning neural network control algorithm are compared and analyzed.The simulation experiment results show that the latter is superior to the previous two methods in control effectiveness and stability,and could effectively improve the control accuracy and robustness of the rudder angle during ship navigation.
关键词
实时反馈/强化学习/神经网络/船舶艏摇Key words
real-time feedback/reinforcement learning/neural network/ship yaw引用本文复制引用
基金项目
山东省船舶控制工程与智能系统工程技术研究中心科研专项(SSCC-2021-0006)
出版年
2024
国家科技期刊平台
NSTL
维普
万方数据