Safety-guaranteed adaptive neural motion control for a hovercraft with multiple constraints

Abstract This paper develops a safety-guaranteed adaptive neural motion controller for an underactuated hovercraft in the presence of multiple constraints and model uncertainties. First, a surge controller is designed based on a new improved integral barrier Lyapunov function (iBLF) to guarantee that the surge speed is above the resistance hump speed to achieve better course stability. Second, the constraint on the drift angle is transformed into one on the sway speed, and then the virtual sway controller is designed to constrain the sway speed for the purpose of confining the drift angle. Third, a time-varying iBLF is constructed to constrain the yaw angular velocity to inside of safety boundary relating to the surge speed to ensure a safe turning motion at high speed. Neural networks (NNs) are incorporated into each controller to handle the model uncertainties. It is proved that all the tracking errors are ultimately uniformly bounded. Finally, the results from numerous simulations demonstrate the effectiveness of the proposed control approaches.