Ship Adaptive RBF Neural Network Course Keeping Control Considering System Uncertainty

Abstract

An adaptive RBF neural network-based nonlinear feedback heading keeping control scheme is proposed for the problem of uncertainty in the dynamic parameters and perturbations of a surface ship's heading keeping model under input saturation. An adaptive neural network technique is used to estimate the model dynamic parameters and external time-varying perturbations, while the minimum learning parameters are used to reduce the computational load, and subsequently, an adaptive neural network nonlinear feedback control scheme is designed using a function with input saturation characteristics embedded in the control law. On the basis of Lyapunov's theorem, it is shown that all signals are consistently bounded in a perturbed uncertain heading-holding system. Finally, the simulation and comparison verify the effectiveness of the designed control scheme.