High resolution methods in presence of noise correlated sensors outputs-Parametric approachs

This paper is devoted to the estimation of noise correlations between sensors. The knowledge of the structure of noise correlations is indeed a basic hypothesis for use of high resolution methods. Three types of methods are presented, using parametric noise modelizations. Furthermore, the numerical studies of the associated problems are presented. These methods give satisfying results on simulations and also on real data and can bring a substantial improvement for practical use of high resolution methods ; furthermore their computation time are not prohibitive for practical applications.