Performance analysis of passive localization strategies: Direct one step approach versus 2 steps approach

In passive emitter localization context the problem of source localization by means of several arrays of sensors (base stations network) traditionally consists in 2 steps. In a first step, intermediate parameters (like angles-of-arrival or times-of-arrival ...) are first estimated separately on each base station. In a second step, the positions are then estimated on a central process unit by means of the transmitted parameters. Recently, studies provide new algorithms dealing directly with all signals of all base stations together and providing the position of the emitters in a centralized "one step" procedure. In this article, we compare both approaches. We show that the differences lie in model structure relaxations, impacting the statistical performances. This study underlines the major improvements available for centralized one step processes.