RBF Neural Network-based PID controller design for the airborne satellite antenna servo system

Both PID method and RBF Neural Network-based PID method are analyzed for the airborne satellite antenna with stabilization and tracing of a two-axis gimbaled tracker. Based on the neural network scheme, a PID controller was used to control the system. While it can not adapt to the variety of disturbance, and its parameters are difficult to obtain. In this paper, RBF neural network parallel with the PID controller was designed. Compared with the conventional controller, RBF neural network parallel with the PID controller can constantly revise the PID parameters by online learning strategy and improve the stabilization and tracing capability of the system.