Neural network-based stochastic adaptive attitude control for generic hypersonic vehicles with full state constraints

Stochastic uncertainties are often encountered and usually regarded as one of the most challenges in the flight control for generic hypersonic vehicles (HSVs). Unfortunately, most of the existing control results still have limitations in handling the problem of stochastic multiple uncertainties. In this paper, we focus on the adaptive control design for the generic HSVs subjected to stochastic multiple uncertainties and full state constraints. By introducing a one to one nonlinear mapping, the HSV system with full state constraints is transformed into a novel nonlinear multivariable system. Additionally, the obstacle caused by unknown time-varying disturbances and stochastic uncertainties can also be effectively circumvented with the fusion of a smooth function and the adaptive bound estimation. Moreover, several Radial basis function (RBF)neural networks (NNs)are used to approximate unknown nonlinear continuous functions. With a stochastic Lyapunov process, all the signals in the closed-loop system are proved to be semi-globally uniformly ultimately bounded and states constraints be finally satisfied. Simulation results have demonstrated a superior performance of the proposed control scheme in comparison to the other state-of-art work.