Adaptive Neural Network Asymptotic Tracking Control for Autonomous Surface Vehicles

Abstract

In this paper, an adaptive neural network asymptotic tracking control method is presented for autonomous surface vehicles (ASV) with unknown uncertainties. In the control design, a smooth function is integrated into backstepping construction, which can achieve asymptotic convergence. The neural networks are borrowed to approximate the lumped nonlinear functions encompassing the unknown dynamics. It can be proved that the tracking errors of ASV can asymptotically converge to zero and all signals of the ASV closed-loop system are bounded. We offer simulation images to exhibit the validity of the devised asymptotic control method of ASV.