Unconditional maximum likelihood approach for near-field source localization

Abstract

Localization of near-field sources requires sophisticated estimation algorithms. In this paper, we propose an unconditional maximum likelihood method for estimating direction of arrival and angle parameters of near-field sources. However, calculation of ML estimation from the corresponding likelihood function results in a difficult nonlinear constraint optimization problem. We therefore employed an expectation/maximization iterative method to obtain ML estimates. The most important feature of the EM algorithm is that it decomposes the observed data into its components and then estimates the parameters of each signal component separately providing a computationally efficient solution to the resulting optimization problem.