Robust iterative learning fault-tolerant control for unsupervised path following of an ASV via the AIS data-based guidance

Considering a trajectory reproducibility mission of an autonomous surface vehicle (ASV), this paper investigates an unsupervised path following control framework that uses a fusion mechanism of the Automatic Identification System (AIS) data and an improved iterative learning control (ILC) strategy. The proposed framework consists of two modules: the guidance term and the control term. For the former, a dynamic virtual ship (DVS) guidance principle is developed by combining the data of AIS, leads to a smooth desired path. Linked with the guidance principle, an improved ILC algorithm is designed that fuses an explicitly fact that an unpredictable actuator fault may be encountered due to the marine disturbance. Note that the tracking accuracy can be increased while the number of iterations is adding. Besides, the dynamic surface control (DSC) and robust neural damping technique are introduced for approximating the model uncertainties and the derivative of the virtual control law. Through the Lyapunov theorem, the semi-global uniform ultimate bounded (SGUUB) properties of the proposed control algorithm are proved. Finally, two numerical examples, including a comparative simulation and a simulated path following operation, are carried out for illustrating the superiority and effectiveness of the proposed framework.