RBF neural network based adaptive sliding mode control for hypersonic flight vehicles

According to hypersonic flight vehicles of highly nonlinear and strong coupling characteristics, RBF neural network (RBFNN) based sliding mode controller is proposed. This aircraft model is multivariable, unstable and assumed to have unknown dynamics. The unknown dynamics is due to the parameters uncertainties. The uncertainty parameters are approximated by RBFNN. Sliding mode control has the flaw of chattering, which may activate the un-modeled dynamics produced by uncertainty parameters. Sliding mode controller with RBFNN by identifying uncertainty parameters can remove the chattering, at the same time, it reserves the fast, robust, immunity of sliding mode control. Simulation studies are conducted for trimmed cruise condition. Simulation for propose scheme are compared with pure sliding mode control, adaptive sliding mode control (ASMC) and linear quadratic regulator (LQR) control. In the existence of certain parameters uncertainty, the proposed scheme not only has quicker tracking speed, but also has well robust performance. Simulation results demonstrate that the performance requirements are met.