基于新型PID神经网络的自适应控制系统研究  

Adaptive Control System Based on a New PID Neural Network

作　　者：吴志敏[1] 

机构地区：[1]江苏工业学院计算机系,常州213016

出　　处：《科学技术与工程》2008年第1期76-79,共4页Science Technology and Engineering

摘　　要：提出一种新的PID型神经网络的自适应控制系统,该控制系统采用对角递归神经网络辨识对象的正向模型,采用一种新型神经网络控制器产生控制量,与常规PID控制不同的是,该控制量不再是误差信号的比例、积分和微分量的简单线性组合,而是这些信号的一种非线性组合,从而可以有效地解决常规PID控制器存在的快速性和超调量之间的矛盾。仿真实验表明,这种新型控制系统具有较强的自适应性和鲁棒性。A new adaptive control system based on a new PID neural network is presented. The control system uses diagonal recurrent neural network to identify the forward model of the plant. The control quantity is produced by a new neural network controller. The difference from the conventional PID is that the control quantity is not the simple linear combination of the proportion, integral and differential coefficient of the error signal, but a nonlinear combination of them. So it solves the problem exiting in conventional PID control. The simulation result shows the control system has strong adaptive ability and robustness.

关 键 词：递归神经网络 自适应控制 PID控制 

分 类 号：TP273.2[自动化与计算机技术—检测技术与自动化装置]