DocumentCode
2684833
Title
Two-channel DOA estimation usign frequency selective music algorithm with a phase compensation in reverberant room
Author
Yang, Jae-Mo ; Choi, Min-Seok ; Kang, Hong-Goo
Author_Institution
Dept. of Electr. & Electron. Eng., Yonsei Univ., Seoul
fYear
2008
fDate
21-23 July 2008
Firstpage
365
Lastpage
368
Abstract
This paper proposes a robust two-channel frequency selective multiple signal classification (MUSIC) method to find a direction of arrival (DoA) information of speech signal. To overcome a phase distortion caused by reverberation and background noise in real acoustic room environments, we adopt a least square (LS)-based phase estimation method. In the phase compensation stage, distorted phases are replaced by estimated phases to enhance the accuracy of covariance matrix needed for the eigen-decomposition of the MUSIC method. Simulation results verify that the proposed algorithm shows much higher estimation accuracy than conventional one while its complexity can be reduced by the frequency selection method.
Keywords
acoustic signal processing; covariance matrices; direction-of-arrival estimation; eigenvalues and eigenfunctions; least mean squares methods; phase estimation; reverberation chambers; signal classification; speech processing; covariance matrix; direction of arrival information; eigen-decomposition; frequency selection method; frequency selective MUSIC algorithm; least square based phase estimation method; multiple signal classification method; reverberant room; speech signal; two-channel DOA estimation; Background noise; Direction of arrival estimation; Frequency estimation; Least squares approximation; Multiple signal classification; Noise robustness; Phase distortion; Phase estimation; Reverberation; Speech; direction of arrival(DoA) estimation; multiple signal classification(MUSIC);
fLanguage
English
Publisher
ieee
Conference_Titel
Sensor Array and Multichannel Signal Processing Workshop, 2008. SAM 2008. 5th IEEE
Conference_Location
Darmstadt
Print_ISBN
978-1-4244-2240-1
Electronic_ISBN
978-1-4244-2241-8
Type
conf
DOI
10.1109/SAM.2008.4606891
Filename
4606891
Link To Document