The research of the direct neural controllerapplied to DC servo motor speed and position control

• Main Authors: Tzeng Gow-Long, 曾國隆
• Other Authors: Chu Ming-Huei
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/64952009279636923991

碩士 === 東南科技大學 === 機電整合研究所 === 97 === The DC servo motors are widely used in industry application. This research applies neural control to improve the stability, adaptability and self-learning ability of control system. The simulations and experiments are executed to investigate the dynamic responses for the control system requirement of high performance and accurate. This research studies the biological neuron structure and human neural network, then study the artificial neural network (ANN), which is applied to execute OR and XOR logical gates mapping problem. The convergent properties are analyzed and simulated with high convergent speed. This study applied ANN controller to DC servo motor speed regulator with load disturbance. The proposed neural controller does not need plant dynamical model, reference model and neural emulator. The simulation and experiment results show that the proposed speed regulator is available to keep motor in constant speed with high convergent speed. This study also applied ANN and ANN＋PD hybrid control to DC servo motor position control. The simulation results show that both of the ANN and hybrid controllers make control system have stable responses. The simulation results also reveals that the dynamic responses will be improved by previous training by a square command signal, which makes the neural network with appropriate initial values.