Robust Neural Network Trajectory-Tracking Control of Underactuated Surface Vehicles Considering Uncertainties and Unmeasurable Velocities

Abstract

This article focuses on the trajectory-tracking of an underactuated surface vehicle (USV) considering model uncertainties and nonlinear environmental disturbances. For trajectory tracking in an actual USV sailing environment, both the inertia and damping matrixes are not diagonal, the velocities states are unmeasurable, and error constraints and input saturation are considered. A robust control strategy is proposed based on the backstepping method, state transformation, a super-twisting state observer, and neural networks. All the closed-loop signals are uniformly ultimately bounded, which is proved by the Lyapunov stability theory analysis. The advantages of the proposed method are as follows. (i) A super-twisting observer is constructed to solve the problem of the velocities being unmeasurable, and the error between the virtual and actual velocities converges to a small neighborhood around zero. (ii) Additional controllers are developed to address input saturation of the system control. (iii) A predefined function design is employed to guarantee the transient trajectory-tracking performance. Finally, simulation results verify the feasibility and effectiveness of the proposed USV trajectory-tracking control method.

Keywords

• Underactuated surface vehicle
• trajectory tracking
• prescribed performance
• neural network
• output feedback control