Based on RBF Neural Network of Hypersonic Re-entry Vehicle Attitude Control

In this paper, an attitude control method combining radial basis (RBF) neural network with integral sliding mode control is proposed for the re-entry stage of hypersonic vehicle with uncertain aerodynamic parameters and atmospheric density. Firstly, the Time-scale separation model of nonlinear equations for aircraft is established. Meanwhile, the feedback linearization method is used to linearize the time scale separation model. For fast and slow control subloops, a global sliding mode variable structure control was designed, and the stability of the closed-loop system was verified by Lyapunov theory. Finally, RBF neural network online regulation law is designed to adjust the controller parameters online to reduce chattering. The simulation results show that the controller can maintain good dynamic characteristics even when the aerodynamic data are greatly deviated.