Experimental Study of Feature Extraction Based on Independent Component Analysis and Evaluation of Feature Independence

曾志强; 王俊元; 马维金

您当前的位置：

首页 >

文章列表页 >

Experimental Study of Feature Extraction Based on Independent Component Analysis and Evaluation of Feature Independence

更新时间：2026-09-14

- Experimental Study of Feature Extraction Based on Independent Component Analysis and Evaluation of Feature Independence
- 基于ICA的特征提取实验研究及特征独立性评价
- Vol. 30, Issue 6, Pages: 524-529(2009)
- 作者机构：
  
  中北大学机械工程与自动化学院
- 作者简介：
- 基金信息：
- DOI：
  CLC：
- Published：2009
- 稿件说明：
移动端阅览
曾志强, 王俊元, 马维金. Experimental Study of Feature Extraction Based on Independent Component Analysis and Evaluation of Feature Independence[J]. 2009, 30(6): 524-529.
DOI：

曾志强, 王俊元, 马维金. Experimental Study of Feature Extraction Based on Independent Component Analysis and Evaluation of Feature Independence[J]. 2009, 30(6): 524-529. DOI：

摘要

在Adams环境下设计了悬臂梁激振仿真系统并在实验室搭建了悬臂梁激振系统

通过对系统多种方式的激振

采集了梁上各测点振动信号.使用MSNR

GOSA和JADE这3种ICA算法分别对信号进行处理;对记录的一组振动信号进行加噪处理;用独立性评价指数衡量了ICA分离模态频率的可靠性.实验结果表明:ICA算法能够有效地提取系统结构模态;3种ICA算法的效果随加噪程度的增大而变差.研究为解决具体工程问题提供了参考.

Abstract

A cantilever excitation simulation system was designed in the Adams environment and a cantilever excitation system was built in laboratory.Vibration signals of measuring points of the cantilever were collected by loading various types of excitation on these two systems.The signals were separately processed by means of three kinds of ICA algorithm of MSNR

GOSA and JADE.Some noise processing was interfused into a group of recorded vibration signals and the independent performance index was used to measure the reliability of the modal frequency separated by ICA.The results show that ICA algorithm can separate the independence source signal effectively

namely it can extraction the system's modal.The effect of three ICA algorithms deteriorate with the increase of noise scale.And this research can provides valuable reference to the practical engineering problems.

关键词

Keywords

references

张小兵,马建仓,陈翠华,等.基于最大信噪比的盲分离算法[J].计算机仿真,2006,23(10):72-75.Zhang Xiaobing,Ma Jiancang,Chen Cuihua,et al.A blind source separation algorithm based on maximum signalnoise ratio[J].Computer Simulation,2006,23(10):72-75.(in Chinese)

王建国,杨万扣,郑宇杰,等.一种基于ICA和模糊LDA的特征提取方法[J].模式识别与人工智能,2008,21(6):819-823.Wang Jianguo,Yang Wankou,Zheng Yujie,et al.A feature extration method based on ICA and fuzzy LDA[J].Pattern Recognition and Artifical Ingelligence,2008,21(6):819-823.(in Chinese)

Lee T W,Girolami M,Bell A J.A unifying information theoretic framework for independent component analysis[J].Computers And Mathematics With Application,2003,31(11):1-21.

Peeters B,Roeck G D,Reference-based stochastic subspace identification for output-only modal analysis[J].Mechanical Systems and Signal Processing,1999(13):855-878.

申丽岩,方滨,沈毅.基于负熵极大的独立分量分析方法[J].中北大学学报,2005(6):396-399.Shen Liyan,Fang Bin,Shen Yi.An independent component analysis algorithm based on maximum negentropy[J].Journal of University of China,2005,26(6):396-399.(in Chinese)

冯旭哲,罗飞路,杨俊,等.利用独立分量分析识别多源通信信号[J].测试技术学报,2008,22(3):256-259.Feng Xuzhe,Luo Feilu,Yang Jun,et al.Identification of multiple source communication signals by usingindependent component analysis[J].Journal of Test and Measurement Technology,2008,22(3):256-259.(inChinese)

方利民,林敏.基于FastICA和神经网络的红酒主要品质参数红外检测[J].光谱学与光谱分析,2009,29(8):2083-2086.Fang Limin,Lin Min.Detection of the main quality indicators in red wine with infrared spectroscopy based on fastICA and neural network[J].Spectroscopy and Spectral Analysis,2009,29(8):2083-2086.(in Chinese)

陆文凯,骆毅,赵波,等.基于独立分量分析的多次波自适应相减技术[J].地球物理学报,2004,47(5):886-891.Lu Wenkai,Luo Yi,Zhao Bo,et al.Adaptive multiple wave subtraction using independent component analysis[J].Chinese Journal of Geogthysics,2004,47(5):886-891.(in Chinese)

谢平,李红亮,黄双峰.一种盲源分离的优先进化自适应遗传算法[J].计算机仿真,2009,26(6):220-223.Xie Ping,Li Hongliang,Huang Shuangfeng.An adaptive genetic algorithm for blind source separation[J].ComputerSimulation,2009,26(6):220-223.(in Chinese)

王俊元.基于ICA的工作模态参数辨识方法研究[D].太原:太原理工大学,2008.

Views

220

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Segmentation Model of Colon Polyp with Lightweight Dynamic Multi⁃Scale

Target Classification Algorithm for SAR Images Based on Multi⁃Attention Feature Fusion

基于R-VGG和多分支注意力的无人机图像配准模型

数字视觉精密测量光纤环绕制状态参数

PEG⁃Like Polymers Based on Low Molecular Weight PEI as Gene Vectors

Related Author

LIU Zesheng

YU Benguo

LIU Zejin

WANG Chunliu

HOU Yuchao

Wang Peng

XU Lilong

雷海卫

Related Institution

School of Software， North University of China

School of Intelligent Medicine and Technology （Big Data Research Center）， Hainan Medical University

School of Mathematics and Computer Science， Shanxi Normal University

School of Mathematics， North University of China

中北大学大数据学院

⁰