Data-driven fault-tolerant path-following control for USV based on fixed-time guidance and fuzzy disturbance observer

This paper investigates the data-driven path-following control of the unmanned surface vessel subject to unknown external disturbances and actuator faults. First, a fixed-time guidance scheme, including a fixed-time sideslip angle observer and a fixed-time line-of-sight guidance law, is proposed to transform the path-following problem into a heading control problem. Next, in the fault-free case, a fuzzy adaptive disturbance observer (FADO)-based model-free adaptive nominal control law is proposed. Further, in the case of unknown time-varying direction faults, neural network is utilized to approximate the bias faults, and an improved Nussbaum function is proposed for handling the fault efficiency factor of unknown time-varying direction, based on which an FADO-based model-free adaptive fault-tolerant control method is proposed. The proposed method is a fully data-driven online learning method that achieves path-following under the constraints of external disturbances and actuator faults solely through input and output data. Finally, the effectiveness and superiority of the proposed method are demonstrated through simulation experiments.