Localization methods for correlated spatially distributed sources

Abstract

Modeling of sources as point in nature is a very idealistic modeling in array processing. There have been a number of distributed source models proposed in the literature. This work is based on one of these models. It tries to extend presently existing algorithms which work well for an uncorrelated source scenario to a more general correlated scenario. The first part of the work addresses the parameter estimation problem in an environment of Coherently Distributed Sources. A brief study of the properties of the array steering matrix is carried out for these sources. A weighted form of spatial smoothing has been proposed which does a parameter estimation in an arbitrarily correlated scenario but with a restriction on the form of the parameters. A heuristic algorithm has been proposed for parameter estimation in an arbitrarily correlated scenario. The second part of the work addresses the parameter estimation problem in an environment of Incoherently Distributed Sources. The ESPRIT-based algorithm for the localization of uncorrelated Incoherently Distributed (ID) sources was extended to a specific correlated ID source scenario.