改进的Elman网络在系统辨识中的应用  

Application of Improved Elman Network in System Identification

作　　者：陈耿彪[1] 陈慧勇[1] 贺尚红[1] 

机构地区：[1]长沙理工大学汽车与机械工程学院,湖南长沙410076

出　　处：《机械工程与自动化》2005年第6期48-49,52,共3页Mechanical Engineering & Automation

摘　　要：介绍了E lm an网络在非线性动力学模型辨识中的应用原理,采用改进的E lm an动态递归网络实现对某平整机液压自动厚控系统(HAGC)的模型辨识。实验表明,E lm an网络利用内部状态反馈来描述系统的非线性动力学行为,提高了学习速度,适合于动态系统的实时辨识。The principle of Elman network to identify nonlinear dynamic systems is introduced. An approach based on improved Elman dynamical recursive neural network to identify a hydraulic AGC dynamic model of temper mill is presented. The experiments show that it uses internal state feedback to describe the nonlinear dynamic rules of system, so that its learning speed is improved and it is suitable for real-time identification of dynamic systems.

关 键 词：系统辨识 ELMAN神经网络 液压压下系统 

分 类 号：TH137.9[机械工程—机械制造及自动化] TP183[自动化与计算机技术—控制理论与控制工程]