船舶航迹自动舵的神经网络控制方法研究

本文提出了一种船舶航迹保持的在线神经网络控制器。该控制器能用舵角控制航迹偏差，它去除了对船舶运动模型的辨识过程，它也不需“教师”或离线训练。为克服船舶过冲引起的航迹偏差等缺点，还设置了适当的训练目标。仿真结果表明该控制器具有很好的控制精度、动态特性和鲁棒性。

A Study on the Neural Network Control Method for Ship Track-keeping Autopilot

This paper presents a neural network approach for ship track-keeping autopilot. The neural controller uses rudder angle to regulate the tracking error,its advantages are free from any dependence on the identification of the mathematical model expressing the dynamics of the ship,and the need for a "teacher" or an off-line training process. The appropriate training targets are set to lessen the tracking error being caused by overshoot and other undesirable characteristics. The simulation results reveal that the controller has fine control accuracy,dynamic character and robustness.