PID神经网络控制器的设计及仿真研究

PID神经网络(PID-NN)是一种新型的前向神经元网络,该隐含层单元分别为比例(P)、积分(I)、微分(D)单元,各层神经元个数、连接方式、连接权初值是按PID控制规律的基本原则确定的。PID神经网络控制器是将神经网络和PID控制规律融为一体,既具有常规PID控制器结构简单、参数物理意义明确之优点,同时又具有神经网络自学习、自适应的功能。本文给出了PID-NN控制器的结构形式,计算公式,从理论上证明了PID-NN的收敛性和稳定性,最后对二阶对象下的系统进行了仿真,证明了PID-NN控制器具有较好的自学习和自适应性。

The design and simulation of PID neural network control

The PID neural network is a new type of feed forward neural network.Its hidden layers are composed of proportional neurons,integral neurons and derivative neurons.In each layer,the number of units,the connective form and the primary value of weights are founded on the basic principles of PID.The PID neural network controller can make both neural network and PID control into an organic whole,which has the merit of any PID controller for its simple construction and definite physical meaning of parameters,and also has the self study and adaptive functions of a neural network.This paper presents the configuration of the PID neural network and reasons out related equations as well as verifies its convergence and stabilization in theory.The computer simulation shows its advantage in self study and adaptability.