Application of Neural Network Minimum Parameter Learning Algorithm in Ship's Heading Tracking Control

Abstract

Ship heading sliding mode tracking control algorithm was investigated based on the minimum parameter learning algorithm of Radial Basis Function (RBF) neural network. RBF neural network has been used to approach the uncertainty function of ship nonlinear motion system and unknown external interference. In consideration that the RBF neural network weights is difficult to adjust quickly, the minimum parameter learning algorithm of RBF neural network was used in this paper to design a single estimated parameter instead of neural network adjustment weights. Finally, by means of Lyapunov stability theory, the tracking control law of ship heading was deduced by RBF Neural network. The above controller's solving speed of an adaptive law is faster than traditional neural network control algorithm with less parameters. By this way, the controller's structure is more simple and therefore easier to be implemented and achieved in application.