Nonsingular direct neural control of air-breathing hypersonic vehicle via back-stepping

This paper investigates the design of nonsingular direct neural control systems for the longitudinal dynamics of a generic air-breathing hypersonic vehicle (AHV). The control objective is to steer the velocity and altitude to track the given commands in the presence of unmatched disturbances. For the velocity subsystem, a neural network (NN) is employed to approximate the developed control law directly where a direct neural controller is designed. For the altitude subsystem that is transformed into strict feedback form, two NNs are used to estimate the virtual and actual control laws derived from back-stepping design. Hence the problem of possible control singularity and “explosion of terms” are avoided. More specially, a nonlinear tracking differentiator is introduced to reconstruct the flight-path angle and angle of attack that are difficult and costly to measure in practice. Numerical simulations are presented to verify the feasibility of the proposed control approach.