High resolution direction finding from rectangular higher order cumulant matrices: The rectangular 2Q-music algorithms

Author

Becker, Hanna ; Chevalier, P. ; Haardt, Martin

Author_Institution

I3S, Univ. de Nice Sophia-Antipolis, Sophia-Antipolis, France

fYear

2014

fDate

4-9 May 2014

Firstpage

2242

Lastpage

2246

Abstract

Recently, the 2q-MUSIC (q ≥ 2) direction finding algorithm has been developed for non-Gaussian sources and square arrangements of the 2qth-order data statistics, to overcome the main limitations of MUSIC and to improve the performance of 4-MUSIC for multiple sources. To further improve the performance of the 2q-MUSIC algorithm, the purpose of this paper is to extend the latter to rectangular arrangements of the data statistics, giving rise to rectangular 2q-MUSIC algorithms. It is shown in particular that rectangular arrangements of the higher order (HO) data statistics allow to optimize the compromise between performance and maximal number of sources to be processed. Besides, it also allows a complexity reduction for a given level of performance. These results, completely new, should open new perspectives for HO array processing.

Keywords

matrix algebra; signal classification; statistical analysis; complexity reduction; data statistics; high resolution direction finding; multichannel signal classification; nonGaussian sources; rectangular 2q-MUSIC algorithm; rectangular higher order cumulant matrices; Acoustics; Conferences; Speech; Speech processing; TV; Yttrium; 2q-MUSIC; Arrangements; Higher order; Rectangular; Virtual Array;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing (ICASSP), 2014 IEEE International Conference on

Conference_Location

Florence

Type

conf

DOI

10.1109/ICASSP.2014.6853998

Filename

6853998