DocumentCode
3102565
Title
Detection criteria and spatial-frequential smoothings
Author
Tas, I. ; Bourennane, S. ; Jourdain, G. ; Faure, B. ; Lacoume, J.L.
Author_Institution
CNRS, St. Martin d´´Heres, France
fYear
1988
fDate
3-5 Aug 1988
Firstpage
135
Lastpage
140
Abstract
Three criteria used to detect how many uncorrelated excitations reach a passive array are examined in two cases: narrowband signals and wideband signals. Two of the criteria are the AIC-MDL (Akaike information-theoretic criterion/minimum description length) criteria and the third uses an adapted statistic of the logarithm of the generalized likelihood ratio. These criteria are compared using analytical simulations, and their performances in real situations are pointed out. The spatial smoothing procedure is also presented. It is shown that, by using both the diagonal elements of the spectral covariance matrix and the spectral smoothed covariance matrix, it is possible to compute the source correlation coefficients. Results on narrow- and wide-band source data recorded during an underwater acoustic experiment are presented
Keywords
acoustic signal processing; information theory; signal detection; statistics; underwater sound; AIC-MDL; Akaike information-theoretic criterion; adapted statistic; detection criteria; generalized likelihood ratio; minimum description length; narrowband signals; passive array; source correlation coefficients; spatial-frequential smoothings; spectral covariance matrix; spectral smoothed covariance matrix; uncorrelated excitations; underwater acoustic experiment; wideband signals; Acoustic sensors; Array signal processing; Covariance matrix; Maximum likelihood estimation; Narrowband; Sensor arrays; Smoothing methods; Testing; Underwater acoustics; Wideband;
fLanguage
English
Publisher
ieee
Conference_Titel
Spectrum Estimation and Modeling, 1988., Fourth Annual ASSP Workshop on
Conference_Location
Minneapolis, MN
Type
conf
DOI
10.1109/SPECT.1988.206178
Filename
206178
Link To Document