A radial basis function Neural Networks based space-vector PWM controller for voltage-fed inverter

In this paper, the basic principle of space-voltage vector PWM (SVPWM) is presented. Due to back propagation neural networks (BP) based SVPWM controller have local optimization problem and lower training rate, a radial basis function neural networks (RBF) controller based SVPWM is proposed. Using Matlab/Simulink together with Neural Network Toolbox, we develop a computer simulation program for the RBF-SVPWM Inverter. The results indicate that the RBF-SVPWM Inverter generates less current harmonic distortion than BP- SVPWM Inverter and the traditional SVPWM Inverter.