Underdetermined blind delayed source separation based on single source intervals in frequency domain

被引：0

作者：

Xiao, Ming ^{[1
,2
]}

Xie, Sheng-Li ^{[1
]}

Fu, Yu-Li ^{[1
]}

机构：

[1] School of Electrics and Information Engineering, South China University of Technology, Guangzhou 510641, China

[2] Department of Electrics and Information Engineering, Maoming College, Maoming 525000, China

来源：

Tien Tzu Hsueh Pao/Acta Electronica Sinica | 2007年 / 35卷 / 12期

关键词：

Acoustic waves - Algorithms - Channel estimation - Computer simulation - Independent component analysis - Mathematical models - Mathematical programming - Reverberation - Sensors - Signal reconstruction;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

How to separate delayed mixtures without any reverberation is studied. An algorithm to estimate channel with delays for the underdetermined model (i.e., fewer sensors than sources) is proposed. It is called attenuation and delay matrix recovery in single source intervals (AD-MRISSI) algorithm. Theoretically, it can estimate the channel without bias. Practically, it can estimate the delays and the attenuations exactly. An algorithm to recover the sources is also proposed based on a sparse criterion. It can avoid solving second-order cone programming problem in Bofill's algorithm. Several experiments of sound sources demonstrate the effectiveness and practicality of our algorithm.

引用

页码：2279 / 2283