论文部分内容阅读
提出一种船舶航迹保持的在线神经网络控制器 .该控制器能解决精确的船舶动态模型难以建立的问题 ,能用舵角同时控制航迹偏差和航向偏差 ,能通过对控制精度的直接计算来自动地在线训练学习而不需离线训练学习过程 .计算机仿真结果表明了该控制器训练方法的有效性和控制的鲁棒性
An online neural network controller for ship track keeping is proposed.The controller can solve the problem that the accurate ship dynamic model is difficult to establish and the rudder angles can be used to control the track deviation and the heading deviation simultaneously.According to the direct calculation of control accuracy To learn online automatically without learning the process of training offline.Computer simulation results show the effectiveness of the controller training method and the robustness of the control