An efficient neural network based tracking controller for autonomous underwater vehicles subject to unknown dynamics

This paper proposes an efficient neural network (NN) controller for the tracking control of an autonomous underwater vehicles (AUV) subject to unknown vehicle dynamics and significant uncertainties. The controller is first designed based on the error dynamics by using backstepping technique. Then, the unknown dynamics and uncertainties of the vehicle are handled by introducing a NN with single-layer structure. The design of the NN is based on the vehicle regressor dynamics that expresses the highly nonlinear dynamics in a linear form in terms of the known and unknown dynamic parameters. The big advantage of the proposed tracking controller is that the learning algorithm of the NN is simple and computationally efficient. In addition, the developed controller is capable of compensating bounded unknown disturbances. The tracking errors are proved to uniformly ultimately bounded and converge to a small neighbourhood of the origin. The effectiveness and efficiency of the proposed controller is demonstrated by simulations results.