基于无监督学习的源数估计及盲分离算法

Journal of Systems Engineering and Electronics ›› 2009, Vol. 31 ›› Issue (8): 1790-1794.

基于无监督学习的源数估计及盲分离算法

谭北海, 赵敏, 谢胜利

华南理工大学电子与信息学院, 广东, 广州, 510640

收稿日期:2008-06-25 修回日期:2008-09-28 出版日期:2009-08-20 发布日期:2010-01-03
作者简介:谭北海(1980- ),男,讲师,博士后,主要研究方向为信号盲分离以及自适应信号处理.E-mail:bhtan@scut.edu.cn
基金资助:
国家自然科学基金重点项目(U0635001);国家自然科学基金(60674033,60774094);中国博士后科学基金(20080430828);广东省自然科学基金(05006508);华南理工大学博士后创新科学基金(20080217)资助课题

Sources number estimation and blind separation algorithm based on unsupervised learning

TAN Bei-hai, ZHAO Min, XIE Sheng-li

School of Electronic and Information Engineering, South China Univ. of Technology, Guangzhou 510640, China

Received:2008-06-25 Revised:2008-09-28 Online:2009-08-20 Published:2010-01-03

摘要/Abstract

摘要： 针对目前欠定盲分离问题中源数未知,采取"两步法"进行分离源信号.在第一步聚类算法中,一般都假设源信号个数已知,即事先给定聚类数目,这类算法成功与否依赖于源信号个数的先验知识.为了有效解决这个问题,提出了一种新的基于模糊聚类分析的无监督学习算法,它利用观测信号之间的相似关系来确定模糊相似矩阵进行迭代算法,不但可以精确估计源信号个数,同时也能获得对混叠矩阵的精确估计.该方法进一步完善了"两步法",仿真结果表明了算法的有效性及优异性能.

Abstract: The two-step approach is often used to separate sources in underdetermined blind separation problem.The first step is to estimate the mixing matrix by clustering algorithms using the observations,in which it is often supposed that the sources number is known,so the two-step approach depends on the prior information about sources number strongly.A novel underdetermined blind separation algorithm based on fuzzy clustering and fuzzy similar matrix is proposed,which can accurately estimate the sources number and the mixing matrix respectively.The last simulations show the good performance of the proposed algorithm.

中图分类号:

TN911.7

谭北海, 赵敏, 谢胜利. 基于无监督学习的源数估计及盲分离算法[J]. Journal of Systems Engineering and Electronics, 2009, 31(8): 1790-1794.

TAN Bei-hai, ZHAO Min, XIE Sheng-li. Sources number estimation and blind separation algorithm based on unsupervised learning[J]. Journal of Systems Engineering and Electronics, 2009, 31(8): 1790-1794.

参考文献

[1] He Z S,Xie S L,Fu Y L.Sparse representation and blind source separation of ill-posed mixtures[J].Science in China Series F-Information Science,2006,49(5):639-652.
[2] Hyvarinen A,Oja E.Independent component analysis:algorithms and applications[J].Neural Networks,2000,13(4-5):411-430.
[3] Hyvarinen A.Fast and robust fixed-point algorithms for independent component analysis[J].IEEE Trans.on Neural Networks,1999,10(3):626-634.
[4] Xie S L,He Z S,Fu Y L.A note on Stone's conjecture of blind separation[J].Neural Computation,2005,17:245-319.
[5] Li Y Q,Amari S,Cichocki A,et al.Underdetermined blind source separation based on sparse representation[J].IEEE Trans.on Signal Processing,2006,54(2):423-437.
[6] Belouehrani A,Cardoso J F.Maximum likelihood source separation for discrete sources[C]//Proc.of EUSIPCO,1994:768-771.
[7] Zibulevsky M,Pearlmutter B.Blind source separation by sparse decomposition in a signal dictionary[J].Neural Computation,2001,13(4):863-882.
[8] Bofill P,Zibulevsky M.Underdetermined source separation using sparse representation[J].Signal Processing,2001,81:2353-2362.
[9] Li Y Q,Cichocki A,Amari S.Analysis of sparse representation and blind source separation[J].Neural Computation,2004,16(6):1193-1234.
[10] Lewicki M S,Sejnowski T J.Learning overcomplete representations[J].Neural Computation,2000,12:337-365.
[11] He Z S,Xie S L,Fu Y L.Sparsity analysis of signals[J].Progressin Natural Science,2006,16(8):879-884.
[12] Lv Q,Zhang X D.A unified method for blind separation of sparse sources with unknown source number[J].IEEE Signal Processing Letters,2006,13:49-51.
[13] 刘海林.基于广义特征值的病态混叠盲源分离算法[J].电子学报,2006,34(11):2072-2075.
[14] Liu H L,Hou J X.A new approach to underdetermined blind source separation using sparse representation[J].LNCS,2007,4481:276-283.
[15] He Z S,Xie S L,Ding S X,et al.Convolutive blind source separation in the frequency domain based on sparse representation[J].IEEE Trans.on Audio Speech and Language Processing,2007,15:1551-1563.
[16] He Z S,Xie S L,Zhang L Q,et al.A note on Lewicki-Sejnowski gradient for learning overcomplete representations[J].Neural Computation,2008,20:636-643.
[17] 张洪渊,贾鹏,史习智.确定盲分离中未知信号源个数的奇异值分解法[J].上海交通大学学报,2001,35(8):1155-1158.
[18] 李广彪,许士敏.基于源数估计的盲源分离[J].系统仿真学报,2006,18(2):485-488.
[19] 张翼,柯亨玉.文必洋,等.相位法估计信号源数[J].武汉大学学报,2003,49(1):137-140.
[20] 章新华,张安清,孙剑平.一种信号源数目的盲估计方法[J].系统工程与电子技术,2001,23(9):9-11.(Zhang Xinhua,Zhang Anqing,Sun Jianping.A blind method for estimating the number of signal sources[J].Systems Engineering and Electronics,2003,23(9):9-11.)
[21] Wong C C,Chen C C.A hybrid clustering and gradient descent approach fuzzy modeling[J].IEEE Trans.on System,Man,and Cybernetics-Part B,1999,29(6):686-693.