模糊神经网络PID在数字舵机控制中的应用

Application of Fuzzy Neural Network PID in Digital Servo Control

摘要:

针对某型号数字舵机系统在非线性时变的复杂条件下,传统的PID控制器响应速度慢,精度低,抗过载能力差的缺点,通过对模糊神经网络算法的研究,结合传统的PID控制器,设计了模糊神经网络PID控制器。分别将两种控制算法应用到舵机系统进行实验可以得出,模糊神经网络PID控制器使得舵机位置环阶跃响应上升时间从80ms减小到35ms,超调量从10％减小到小于5％,在频率特性测试中反馈曲线衰减从-2.79dB减小到-0.77dB,相移从65°减小到35°,同时系统非线性引起的畸变明显改善。

Abstract:

For the nonlinear and time variation characteristics of a certain type of servo, the traditional PID algorithm has a shortage of slow-response, low-accuracy and the ability of anti-overload is poor. Acorrding to the study of Fuzzy Neural Network algorithm, combine with the traditional PID controller, the Fuzzy Neural Network PID controller is designed. By applying the two algorithms to the servo control system and taking experiments, the result shows that the Fuzzy Neural Network PID controller makes the rise time of the step response reduces from 80ms to 35ms, the overshoot decreases from 10％ to less than 5％, then in the frequency response test, the attenuation of feedback curve reduces from -2.79dB to -0.77dB and the phase-drift decreases from 65° to 35°, further more, the distortion caused by the system nonlinear improve clearly.