SAR Power Controller Design Research Based On PID Neural Network Artificial Intelligence and CPSO Algorithm

Abstract

The $S^{4}R$ Switch topology technology is used in SAR satellite power system which has nonlinear characteristics, and difficult to establish an accurate mathematical model for designing a satisfied controller. In this paper, developed an artificial intelligence design method based the Dynamic P-I-D Neurons Identification Model (DPIDNIM) and nonlinear auto regressive moving average (NARMA-L2) model. In order to verify the design method’s feasibility, used a typical non-linear time series named Mackey-Glass (MG) time series to confirm the accuracy and effective of the identification model. Then a cooperated-PSO-based PID neural network control strategy (PIDNNC) is given out. Research results indicate DPIDNIM-CPSO-Based Artificial Intelligence design method adapts to design a controller for SAR satellite power system