Cost and other practically related reasons can mean that velocity sensors are not available on an underwater vehicle. For such cases, the results in this brief are developed on an observer-based fault-tolerant control for underwater vehicles in the presence of external disturbances and unknown thruster faults. An adaptive sliding mode observer is developed to achieve finite-time convergence where, in comparison to a high-gain-based design for the observer, a nonlinear feedback is constructed based on the position estimation error. Unlike alternatives, a discontinuity term in the developed fault tolerant controller is avoided, and the stability of the controlled dynamics is characterized using the Lyapunov theory. Finally, these new results are supported by both a simulation-based study and experimental verification.

A region tracking fault tolerant control approach based on backstepping technique is proposed for Autonomous Underwater Vehicles (AUV). The proposed approach aims at driving tracking error to reach into the desired region in presence of ocean current disturbance, modelling uncertainty, unknown thruster faults and thruster amplitude and rate saturation constraints. At first, a type of piecewise and differential Lyapunov function is constructed to achieve region tracking control in the frame of backstepping technique. Then, the paper analyses and acquires the bound structures of the lumped uncertainty (including ocean current disturbance and model uncertainty) and the variation of thruster distribution matrix caused by unknown thruster faults, respectively. An adaptive technique is used to estimate the unknown coefficients in the above bound structures. In addition, an adaptive adjustment scheme for the desired trajectory is developed to achieve region tracking control with thruster amplitude and rate saturation constraints. The stability of the closed-loop system is analysed based on Barbalat's lemma. Finally, simulations and pool-experiments are presented to illustrate the effectiveness of the proposed method.