Cet article introduit, dans le contexte de la localisation de sources, une nouvelle technique d'estimation du Propagateur a partir des statistiques d'ordre quatre. Le Propagateur est un op erateur lin eaire qui d epend des param etres de propagation et qui permet la d etermination du sous-espace bruit sans aucune d ecomposition propre de la matrice interspec-trale des signaux ree cus. A la dii erence des techniques clas-siques d'estimation du Propagateur a partir des statistiques d'ordre deux, cette technique fournit un estimateur asymp-totiquement non biais e. Les performances asymptotiques de la m ethode propos ee ont et e d evelopp ees. Des simulations num eriques illustrent la validit e de la m ethode dans des con-textes diiciles. Abstract This contribution introduces, in the source localization context , a new estimation technique of the so-called Propagator operator from the fourth order statistics. The Propagator is a linear operator depending on the propagation parameters. It allows the noise-subspace determination without any eigen-decomposition of the cross-spectral matrix of the received signals. In contrast to second order estimation methods of the Propagator operator, this method gives an asymptoti-cally unbiased estimator. The performance of this method is assessed by deriving the asymptotic distribution of the DOA (i.e direction-of-arrivals) estimators. Several numerical simulations are presented to demonstrate the eeectiveness of the method.
[1]
J. Cardoso.
HIGHER-ORDER NARROW-BAND ARRAY PROCESSING
,
1991
.
[2]
S. Marcos,et al.
Source localization using a distorted antenna
,
1989,
International Conference on Acoustics, Speech, and Signal Processing,.
[3]
Messaoud Benidir,et al.
Source-bearing estimation and sensor positioning with the Propagator method
,
1990
.
[4]
Eric Moulines,et al.
SECOND-ORDER VERSUS FOURTH-ORDER MUSIC ALGORITHMS : AN ASYMPTOTICAL STATISTICAL ANALYSIS
,
1991
.
[5]
Eric Moulines,et al.
Asymptotic performance analysis of direction-finding algorithms based on fourth-order cumulants
,
1995,
IEEE Trans. Signal Process..
[6]
P. McCullagh.
Tensor Methods in Statistics
,
1987
.
[7]
Munier.
1 - L'identification de fronts d'ondes corrélés et distordus
,
1987
.