基于ICA的雷达信号欠定盲分离算法

陈晓军; 成昊; 唐斌

doi:10.3724/SP.J.1146.2009.00291

基于ICA的雷达信号欠定盲分离算法

doi: 10.3724/SP.J.1146.2009.00291

计量
- 文章访问数: 4241
- HTML全文浏览量: 143
- PDF下载量: 1202
- 被引次数: 0
出版历程
- 收稿日期: 2009-03-09
- 修回日期: 2009-12-07
- 刊出日期: 2010-04-19

Underdetermined Blind Radar Signal Separation Based on ICA

摘要

摘要: 该文针对源信号时域和频域不充分稀疏的情况，提出了欠定盲源分离中估计混合矩阵的一种新方法。该方法对等间隔分段的观测信号应用独立分量分析(ICA)的盲分离算法获得多个子混合矩阵，然后对其分选剔除了不属于原混合矩阵的元素，最后利用C均值聚类的学习算法获得对混合矩阵的精确估计，解决了源信号在时域和频域不充分稀疏的情况下准确估计混合矩阵的问题。在估计出混合矩阵的基础上，利用基于稀疏分解的统计量算法分离出源信号。由仿真结果，以及与传统的K均值聚类，时域检索平均算法对比的实验结果说明了该文算法的有效性和鲁棒性。
- 信号处理; 欠定盲源分离; 独立分量分析; C均值聚类;稀疏分解的统计量
Abstract: A method of the mixing matrix estimation in the underdetermined source separation is proposed in which the sources are not sparse enough to estimate the mixing matrix. Getting many sub matrixes through applying Independent component analysis(ICA) for observation signals and removing the elements do not belong to the mixing matrix, the mixing matrix is estimated precisely with C-means clustering agglomeration. Then, the source signals can be recovered with the statistically sparse decomposition principle. The experiment shows that the method have better accuracy and validity than K-means and searching-and-averaging method in the time domain in estimating the mixing matrix.

HTML全文

参考文献(1)

Aissa-El-Bey, Linh-Trung N, and Abed-Meraim K, et al..Underdetermined blind separation of nondisjoint sources inthe time-frequency domain[J].IEEE Transactions on SignalProcessing.2007, 55(3):897-907[2]Delorme A, Sejnowski T, and Makeig S. Enhanced detectionof artifacts in EEG data using high-order statistics andindependent component analysis[J].Neuroimage.2007, 34(4):1443-1449[3]Bofill P and Zibulevsky M. Underdetermined blind sourceseparation using sparse representations [J]. SignalProcessing, 2001, 81(11): 2353-2362.[4]Donoho D L and Elad M. Optamally sparse representation ingeneral dictionaries via l1-norm minimization[J].Proceedingsof the National Academy of Sciences of the USA.2003, 100(4):2197-2202[5]Li Y Q, Cichoci A, and Amari S. Analysis of sparserepresentation and blind source separation[J]. NeuralComputation, 2004, 16(6): 1193-1234.[6]Fvotte C and Godsill S. A Bayesian approach for blindseparation of sparse sources[J]. IEEE Transactions on AudioSpeech Language Process, 2006, 16(6): 2174-2188.[7]肖明, 谢胜利, 傅予力. 欠定情形下语音信号盲分离的时域检索平均法. 中国科学(E辑), 2007, 37(12): 1564-1575.Xiao M, Xie S L, and Fu Y L. Searching-and-averagingmethod of underdetermined blind speech signal separation intime domain[J]. Sci China(E index), 2007, 37(12): 1564-1575.[8]Araki S, Sawada H, and Mukai R. Underdetermined blindsparse source separation for arbitrarily arranged multiplesensors[J].Signal Processing.2007, 87(8):1833-1847[9]谭北海, 谢胜利. 基于源信号数目估计的欠定盲分离[J].电子与信息学报.2008, 30(4):863-867浏览[10]Hamada T, Nakano K, and Ichijo A. Wavelet-basedunderdetermined blind source separation of speech mixtures.IEEE ICCAS, Seoul, Korea, 2007: 2790-2794.[11]Peng D and Xiang Y. Underdetermined blind sourceseparation based on relaxed sparsity condition of sources[J].IEEE Transactions Signal Processing.2009, 57(2):809-814[12]Cruces S, Castedo L, and Cichocki A. Robust blind sourceseparation algorithms using cumulants[J]. Neural Computing,2002, 49(12): 87-118.[13]Xiao M, Xie S L, and Fu Y L. A statistically sparsedecomposition principle for underdetermined blind sourceseparation. Intelligent Signal Processing andCommunication Systems, Guang Zhou, China, 2005:165-168.

施引文献

资源附件(0)

访问统计