Full-Drive Ship Trajectory Tracking MLP Learning Control

An adaptive iterative sliding mode control system, integrated within a Multi Layer Perceptron (MLP) neural network, is proposed to tackle the trajectory tracking control challenges faced by all-drive ships while accounting for model uncertainties and external environmental perturbations. Initially, a nonlinear disturbance observer is developed to estimate unknown external disturbances. Subsequently, the robust approximation capabilities of a Radial Basis Function (RBF) neural network are employed to address these model uncertainties. A recursive sliding manifold is then formulated by integrating the interrelationships between the ship's positional and velocity errors, supplemented by an adaptive law aimed at mitigating the neural network's approximation errors. Furthermore, the incorporation of the MLP neural network enhances the learning process of the controller, thereby optimizing control performance. Finally, the Lyapunov function is employed to demonstrate the stability and boundedness of the proposed controller within the closed-loop dynamics. The designed controller is evaluated through simulation on a supply vessel, yielding results that highlight the superiority and efficacy of the proposed control strategy.