| 基于对角回归神经网络的自整定全系数自适应控制器及其特征参量辨识 |
| |
| 引用本文： | 王丽娇.基于对角回归神经网络的自整定全系数自适应控制器及其特征参量辨识[J].航天控制,2011,29(5):15-21. |
| |
| 作者姓名： | 王丽娇 |
| |
| 作者单位： | 北京控制工程研究所,北京 100190;空间智能控制技术重点实验室,北京 100190 |
| |
| 摘 要： | 针对全系数自适应控制器参数人为调试带来的不便,提出一种基于对角回归神经网络的参数自整定方法,通过神经网络的自学习能力对全系数自适应控制器参数进行在线整定.同时,提出一种新的特征模型参数间接辨识方法,采用神经网络的权值和回归层输出组成的非线性函数构造对象的特征参量,更有效地对特征模型的时变参数进行自学习和调整.对闭环回路...
|
| 关 键 词： | 全系数自适应控制器 对角回归神经网络 自整定 特征模型 辨识 |
DRNN Based Self-Adjusting of All-Coefficients Adaptive Controller and Identification of Its Characteristics Parameters |
| |
| WANG Lijiao.DRNN Based Self-Adjusting of All-Coefficients Adaptive Controller and Identification of Its Characteristics Parameters[J].Aerospace Control,2011,29(5):15-21. |
| |
| Authors: | WANG Lijiao |
| |
| Institution: | WANG Lijiao1,2 1.Beijing Institute of Control Engineering,Beijing 100190,China 2.Science and Technology on Space Intelligent Control Laboratory,China |
| |
| Abstract: | A self-adjusting method based on DRNN is proposed to solve the inconvenience of the adjusting of all-coefficients adaptive controller.The controller can change its parameters on line automatically for the self-learning ability of the neural network.In addition,a new identification algorithm of the characteristics model parameters is presented.In the algorithm,the nonlinear function derived from the weight and output of the recurrent layer of DRNN is used to establish the characteristics parameters.The chara... |
| |
| Keywords: | All-coefficients adaptive controller DRNN Self-adjusting Characteristics model Identification |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
