Stochastic Adaptive Neural Network Control for Hypersonic Vehicles

The stochastic disturbance is one of the most significant challenges in the development of hypersonic vehicles. Introducing the stochastic system is a potential solution that could solve stochastic disturbance problems. However, the presence of stochastic disturbances renders conventional stability criteria inapplicable for analyzing the stability of nonlinear stochastic systems. To address this challenge, an original criterion of the stability is constructed through the application of the integral mean value theorem. This paper proposes adaptive neural control for hypersonic vehicles with the uncertainty and circumstances of external disturbance. The backstepping methodology is employed to synthesize the target controller, while the RBF neural networks are employed to approximate the unknown nonlinear dynamics. Simulation results demonstrate that the proposed tracking control scheme exhibits enhanced robustness against aerodynamic uncertainties in hypersonic vehicles. Benchmark comparisons with classical deterministic control frameworks reveal superior performance in convergence speed, trajectory tracking precision, and actuation smoothness.