PDF全文下载:2017060001
丁锋1,2, 徐玲2, 刘喜梅1
(1.青岛科技大学 自动化与电子工程学院,山东 青岛 266042;2.江南大学 物联网工程学院,江苏 无锡 214122)
摘要: 递推辨识和迭代辨识构成了两大类辨识方法族。利用递阶辨识原理、多新息辨识理论,针对多频标准正弦信号的建模问题,提出了递阶梯度迭代参数估计方法、递阶多新息梯度迭代参数估计方法、递阶牛顿迭代参数估计方法等。给出了几个典型辨识算法的计算流程和计算步骤。文中的方法可以推广到其它多频信号模型的参数辨识。
关键词: 信号建模; 参数估计; 梯度搜索; 牛顿搜索; 多新息辨识理论; 递阶辨识; 迭代辨识; 正弦信号
中图分类号: TP 273文献标志码: A
引用格式:丁锋, 徐玲, 刘喜梅. 信号建模(6):多频信号模型的递阶迭代参数估计\[J\]. 青岛科技大学学报(自然科学版), 2017, 38(6): 1-13.
DING Feng, XU Ling, LIU Ximei. Signal modeling. Part F: Hierarchical iterative parameter estimation for multifrequency signal models\[J\]. Journal of Qingdao University of Science and Technology(Natural Science Edition), 2017, 38(6): 1-13.
Signal Modeling. Part F: Hierarchical Iterative Parameter
Estimation for Multi-Frequency Signal Models
DING Feng1,2, XU Ling2, LIU Ximei1
(1.College of Automation and Electronic Engineering, Qingdao University of Science and Technology, Qingdao 266042, China;2.School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, China)
Abstract: Recursive identification and iterative identification form two large families of identification methods. By means of the hierarchical identification principle and the multi-innovation identification theory, this paper studies the multi-frequency standard sine signal modeling, and presents the hierarchical gradient based iterative(GI) parameter estimation method and the hierarchical multi-innovation GI parameter estimation method, and the hierarchical Newton iterative(NI) parameter estimation method. Finally, the computational steps and flowcharts of several typical identification algorithms are given. The methods in the paper can be extended to other multi-frequency signal modeling.
Key words: signal modeling; parameter estimation; gradient search; Newton search; multi-innovation identification theory; hierarchical identification; iterative identification; sine signal
收稿日期: 2017-10-25
基金项目: 国家自然科学基金项目(61472195).
作者简介: 丁锋(1963—),男,博士,“泰山学者”特聘教授,博士生导师.
文章编号: 1672-6987(2017)06-0001-13; DOI: 10.16351/j.1672-6987.2017.06.001