Neural Network–Based Efficient Hybrid Control Scheme for the Tracking Control of Autonomous Underwater Vehicles

Abstract

This work proposes the trajectory tracking control problem of an autonomous underwater vehicle subject to parameter uncertainties and unmodelled dynamics. The developed schemes captivatingly consume the advantages of the model-reliant, model-free control scheme, and conventional backstepping control schemes. The available dynamics of the manipulator's system have been used to make a high-quality benefit significantly. The neural network renovation error and the consequences of the disturbance term are remunerated through an adaptive compensator. The system stability is established through Lyapunov stability criteria as well as the Barbalat lemma and tracking errors are shown to be bounded as well as converge to in the neighbourhood of zero asymptotically. Finally, some comparative simulation cases for a 4-DOF autonomous underwater vehicle are established to show the potential of the developed control proposal.