Direct neural discrete control of hypersonic flight vehicle

This paper investigates the discrete neural control for flight path angle and velocity of a generic hypersonic flight vehicle (HFV). First, strict-feedback form is set up for the attitude subsystem considering flight path angle, pitch angle, and pitch rate by altitude-flight path angle transformation. Secondly, the direct Neural Network (NN) control is proposed for attitude subsystem via back-stepping scheme. The direct design is employed for system uncertainty approximation with less online tuned NN parameters and there is no need to know the information of the upper bound of control gain during the controller design. Thirdly, with error feedback and NN design, the semiglobal uniform ultimate boundedness (SGUUB) stability is guaranteed of the closed-loop system. Similar NN control is applied on velocity subsystem. Finally, the feasibility of the proposed controller is verified by a simulation example.