Robust adaptive NN course controller design based on DSC

Abstract

This paper describes an adaptive course controller with rudder dynamics. The model is described by a third order nonlinear model with unknown parameters. An adaptive neural network (NN )control algorithm based on dynamic surface control (DSC) is developed. With only one learning parameter and reduced computation load, the proposed algorithm can avoid both problem of “explosion of complexity” in the conventional backstepping method and singularity problem. In addition, the boundedness stability of the closed-loop system is guaranteed and tracking error can be made arbitrary small. The effectiveness of the presented autopilot has been demonstrated in the simulation.