Bayesian histogram filter localizer performance in a multiple contact, continental shelf environment.

Abstract

Model-based recursive Bayesian state estimation provides a straight forward way of including a priori environmental knowledge in a localization framework. Previous work by Jemmott et al. demonstrates successful range and depth localization of an acoustic source using this method and data from a single hydrophone. However, that work utilized a relatively benign environment from the SWellEx 96 experiment in which only two source signals were present. This talk will extend the work of Jemmott et al. and discuss the performance of the localization scheme in a multiple contact, continental shelf environment. [Work supported by ONR Undersea Signal Processing.]