基于神经网络的IMC-PID超声波电机控制

由于常规PID控制器无法有效地控制超声波电机非线性的运行特性,提出了一种基于神经网络的可变增益型内模PID控制(IMC-PID),构建IMC-PID简化PID控制器的参数整定,只需调节其中一个参数,为弥补超声波电机的非线性问题,引入神经网络到控制器中来调节参数。实验基于超声波电机伺服系统,利用该控制系统对行波超声波电机进行控制,得到了控制输入和输出的响应关系,并得出了较小的稳态误差,实验证明对超声波电机特性变化及负载扰动适应能力强。

IMC-PID Controls Based on Neural Network for Ultrasonic Motor

A variable gain internal model control PID (IMC-PID) based on neural network is proposed for better ultrasonic motor control.The IMC-PID simplifies the parameters setting of PID controller with only one parameter needed to be adjusted.The neural network is introduced to adjust the parameters to make up for the nonlinear problem for ultrasonic motor.The response relationship between input and output and the smaller steady-state error is found.The experiments proved that ultrasonic motor control by using this control system offers good adaptability to characteristics variation and load disturbance.