RBF neural network-based sliding mode control for a ballistic missile

In this paper, the non-linear models for the three channels of a ballistic missile are analysed. The coupled terms are taken as additional disturbances for every single channel, in order to realise the independent design for every channel and to simplify the structure of the control system. An RBF neural network-based sliding mode controller is designed for every channel's thrust vector control system of the ballistic missile. For the controller, the RBF neural network modifies the parameter of the sliding mode controller to approximate the lumped uncertainty. The performance of the designed RBF neural network-based sliding mode controller is compared with that of the conventional PID controller in the numerical simulation, and its effectiveness is demonstrated by the simulation results.