Direct neural control of hypersonic flight vehicles with prediction model in discrete time

In this paper, the direct adaptive neural controller is investigated for the longitudinal dynamics of a generic hypersonic flight vehicle (HFV). The objective of the controller is to make the altitude and velocity to follow a given desired trajectory in the presence of aerodynamic uncertainties. Based on the functional decomposition, the adaptive discrete-time nonlinear controllers are developed using feedback linearization and neural approximation for the two subsystems. Different from the back-stepping design, the altitude subsystem is transformed into the explicit four-step ahead prediction model. With the prediction model, the controller is proposed without virtual controller design. Furthermore, only one direct neural network (NN) is employed for the lumped system uncertainty approximation. The controller is considerably simpler than the ones based on back-stepping scheme and the algorithm needs less NN parameters to be adjusted online. The semiglobal uniform ultimate boundedness (SGUUB) stability is investigated by the discrete-time Lyapunov analysis and the output tracking error is made within a neighborhood of zero. Accordingly, the NN controller is designed for velocity subsystem. The simulation is presented to show the effectiveness of the proposed control approach.