Abstract
In the power supply, stability and accuracy are important indicators. Unstable and inaccurate power supply could damage to the circuit and entire system. For the power supply system, the most common control method is PID control to resolve the issue but it cease to be effective in the nonlinear system and could be affected by noise easily. This paper introduces a PID control algorithm optimized by PSO-BP algorithm and Kalman filter, which introduces neural network to solve the problem that the performance of traditional PID algorithm may be reduced in nonlinear system. PID system parameters are adjusted by adaptive adjusting network weights. At the same time, this paper introduces PSO algorithm. By this way, the initial weight of the network could be optimized. Filtering the input signal by Kalman filter can effectively reduce the impact of noise. This optimized algorithm can effectively improve the unstable output caused by the system process noise and improve the output voltage of PID controller stability in nonlinear system.