Neural network-based adaptive dynamic surface control for an airbreathing hypersonic vehicle

This paper addresses the design of flight control system related to neural network-based adaptive dynamic surface control for the longitudinal motion of an airbreathing hypersonic vehicle. The control objective is to provide adaptive velocity and altitude tracking in the presence of the model uncertainties and unknown nonlinearities caused by changes of flight conditions. By approximating the unknown nonlinear functions by radial basis function networks, we incorporate the dynamic surface technique into a neural network based adaptive control design framework. The framework is adopted to design dynamic state-feedback controllers that provide stable tracking of velocity and altitude subsystems. Stability analysis shows that the control law can guarantee the uniformly ultimate boundedness of the solution of the closed-loop system and make the tracking error arbitrarily small.