压缩感知原理在盲信号分离中的应用.pdfVIP

Vo l. 29 No.3 广东工业大学学报 第 29 卷第3 期 2012 年9 月 Journal of Guangdong University of Technology September 2012 doi: 10. 3969/j. issn. 1∞7-7162.2012.03.009 压缩感知原理在盲信号分离中的应用 王涛文 (广东工业大学应用数学学院，广东广州I 510520) 摘要:主要阐述了压缩感知的基本原理，介绍了压缩感知的 3 个基本问题:信号的稀疏表示、稀疏基与测量矩阵的 不相关性和信号的重构，分析了它与盲信号分离之间的联系，为解决盲信号分离问题提供了一个新的途径.最后通 过具体实验说明它在盲信号分离上的应用. 关键词:压缩感知;稀疏;不相关;信号分离 中图分类号: TN912. 3 文章编号: 1 ∞7-7162(2012 )侣，ω49-05 文献标志码:A The Application of Compressed Sensing in Blind Source Separation Wang Tao-wen (School of Applied Mathematics , Guangdong University of Technology , Guangzhou 510520 , China) Abstract: Compressed Sensing has been a new signal sampling theory in recent years , for it overcomes the high rate of samplìng defects of traditional Nyquist signal sampling theory , It presented the basic prin- ciples of Compressed Sensing , introduced three fundamental questions of Compressed Sensing-the sparse- ness of sìgnals 、 irrelevance between sparse matrix and measurement matrix , and reconstruction of the sig- nals ,and analyzed the contact between Compressed Sensing and Blind Source Separation. Then , it of- fered a new way to solve the problem of Blind Source Separation. Fìnally , through the experiment it showed its application in Blind Source Separation. Key words: compressed sensing; sparse; irrelevance; signal separation 压缩感知理论自 2006 年由 Donoho 与 Candes 多年来，信号的采样一直遵循著名的 Nyquist 信 号采样定理，但 Nyquist 采样定理要求采样速率大于 等人正式提出，就受到信息论、信号处理、图象处理、 光学、