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

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