欠驱动船舶路径跟踪的神经网络稳定自适应控制

Stable adaptive neural network control of path following for underactuated ships

针对三自由度欠驱动船舶模型参数不确定的路径跟踪问题,设计了神经网络稳定自适应控制器.首先应用微分同胚等效变换和Lyapunov直接法设计参考航向和参考速度,然后对船舶模型参数不确定的操纵环路和推进环路分别设计神经网络稳定自适应控制器跟踪参考航向和参考速度,并对外界风浪流干扰进行自适应补偿.Lyapunov稳定性分析证明了船舶路径跟踪闭环系统的所有误差信号一致最终有界.仿真研究验证了神经网络稳定自适应路径跟踪器的有效性.

We propose a stable adaptive-neural-network control for the path following in a 3 degrees-of-freedom （3DOF） underactuated ship with parameters uncertainties and disturbances. Based on the diffeomorphism transformation and Lyapunov direct method, controllers for the reference yaw angle and the reference speed are designed. Signal errors in the closed-loop path following system are proved to be uniformly ultimately bounded in a small neighborhood of zero. Numerical simulation results are given to illustrate the effectiveness of the stable adaptive-neural-network algorithm.