基于Hopfield网络的PID在直流伺服电机中的应用

PID based on Hopfield neural network in the application of DC servo motor

针对直流伺服电机的非线性和时变性因素,本文结合传统PID控制器特点,介绍了一种基于Hopfield神经网络PID控制方法。该方法利用Hopfield神经网络的自学习能力,经过有限次的训练可以得到了PID控制器所需要的最优参数。采用Matlab软件对构造的系统模型进行了仿真和跟踪实验。实验表明这种方法既简化了经典控制PID参数整定,同时使系统具较好的实时性、稳定性和跟踪性,控制效果比较理想。

For the nonlinear and time-varying factors of DC servo motor,we introduced a control method based on Hopfield neural network,combining with features of the traditional PID controller,in this paper.This method used of self-learning ability of Hopfield neural network.After training in the limited time,we can get the optimal parameters for PID controller.Simulated constructed model and did tracking experiment through Matlab.Experiments show that this method not only simplified the tuning complexity of classic PID control parameters,but also enabled real-time,better stability and tracking property.The effect on the control of the controlled object is an ideal.