Adaptive neural network–based trajectory tracking non-affine control of hovercraft with double-loop constraint performance

Abstract

This study investigates the design of a trajectory tracking constraint control method for hovercraft non-affine formulations via neural network approximation. The tracking scheme comprises an outer closed-loop of attitude and an inner closed-loop of velocity (double-loop). First, a performance function is devised to constrain the attitude errors and then a transformed error function is introduced to simplify it into an equivalent unconstrained control system. Next, an improved asymmetric integral barrier Lyapunov function is designed to address asymmetric velocity constraint problems, which can guarantee the safe turning motion or performance required at high speed. Furthermore, neural networks are employed to estimate the unknown terms of each subsystem, which solve system uncertainties and non-affine dynamics problems. The double closed-loop control system is ultimately uniformly bounded by utilizing the Lyapunov stability principle. Finally, simulation results are presented to verify the tracking performance and effectiveness of the proposed control strategy.