Title :
Localization of wideband sources in colored noise VIA generalized least squares (GLS)
Author :
Velni, JavadMohammadpour ; Khorasani, Kash
Author_Institution :
Dept. of Mech. Eng., Houston Univ., TX
Abstract :
In this paper we present a new method for localization of wideband sources in the presence of colored noise. The optimization criterion used is an extension of conventional least-squares method known as generalized least squares (GLS). The crux of the criterion is based on covariance matching and involves its transformation that yields a more efficient and consistent estimate of the direction of arrivals (DOAs) as compared to the LS and GLS methods that are only developed in the literature for narrowband sources
Keywords :
covariance analysis; direction-of-arrival estimation; least squares approximations; optimisation; signal processing; signal sources; DOA estimation; GLS; colored noise; covariance matching; direction of arrival; generalized least square; optimization criterion; wideband source localization; Acoustic sensors; Array signal processing; Colored noise; Direction of arrival estimation; Frequency estimation; Least squares methods; Narrowband; Sensor arrays; Signal processing; Wideband;
Conference_Titel :
Statistical Signal Processing, 2005 IEEE/SP 13th Workshop on
Conference_Location :
Novosibirsk
Print_ISBN :
0-7803-9403-8
DOI :
10.1109/SSP.2005.1628651
