Design of Single Neuron Adaptive PID Based on Quadratic Performance Index for Linear DC Motor

Abstract

In this paper, a new control strategy is put forward, which is a single neuron PID control based on the quadratic performance index (QPI) learning algorithm. This single neuron PID control is self-learning and self-adjusting. Through the study of system performance, the weights of the neuron are adjusted by the QPI learning algorithm. The new strategy solves the contradiction between fast tracking performance and robustness in the positioning servo system. In addition, the mathematic model of voice coil motor is analyzed. In order to improve the fast response performance and disturbance suppression of the position loop, the Two-Degree-of-Freedom (TDOF) control is proposed. The simulation results show that disturbance rejection and dynamic performance of the VCM servo control system have been improved. Compared with the utilizing traditional PID system, this new control strategy has stronger self-adapting and robustness.