Improved finite-time prescribed performance super-twisting sliding mode united control framework for underwater vehicle-manipulator system

Abstract

This paper investigates the trajectory tracking control problem for underwater vehicle-manipulator system (UVMS) with model uncertainties, dynamic coupling effects and external disturbances. Firstly, an improved finite-time performance function (IFTPF) is proposed to ensure the prescribed tracking accuracy with predefined convergence time while avoiding singularities in the system and enhancing robustness. Additionally, to reduce the controller’s dependence on the system model and compensate for unknown external disturbances, a neural network (NN) estimator is proposed to estimate the combination of system uncertainties and dynamic coupling effects, while a neural network-based nonlinear disturbance observer (NNDO) addresses the unknown external disturbances. Furthermore, an improved finite-time prescribed performance super-twisting sliding mode control framework using neural network-based disturbance observer (NNDO-IFTPPSTSMC) is proposed to ensure control accuracy and robustness while reducing chattering phenomenon. Theoretical analysis demonstrates that corresponding error variables of the system are bounded, and the trajectory tracking errors satisfy the prescribed performance constraints. Finally, simulation results validate the effectiveness and feasibility of the proposed method.