Closest Source Selection Using IVA and Characteristics of Mixing Channel

Abstract

This paper introduces a method for selecting a target source of interest. The target source is assumed to be the closest to sensors among all the other sources regardless of the target source not being the dominant power at the sensors. In this paper, we propose a simple method to select the closest source from signals separated by Independent Vector Analysis (IVA). The proposed method is processed in two-stages. Firstly, IVA is used to separate the mixed signals. Secondly, the mixing channel characteristics are used to choose the closest source. Simulated experimental results are presented to show how well the proposed method works. ? 2009 Springer-Verlag Berlin Heidelberg.