Three-dimensional source tracking in an uncertain environment via Bayesian marginalization

Abstract

This paper develops a non-linear Bayesian marginalization approach for three-dimensional source tracking in shallow water with uncertain environmental properties. The algorithm integrates the posterior probability density via a combination of Metropolis-Hastings sampling over environmental and bearing model parameters and Gibbs sampling over source range/depth, with track constraints on source velocity applied. Marginal distributions for source range/depth and source bearing are derived, with source position uncertainties estimated from the distributions. The Viterbi algorithm is applied to obtain the most probable three-dimensional track. The approach is applied to experimental narrowband data recorded on a bottom-moored horizontal line array in the Barents Sea.

Bayesian three-dimensional reconstruction of toothed whale trajectories: Passive acoustics assisted with visual and tagging measurements

“…1-457). Bayesian modeling has already proven to be an efficient framework to address advanced issues in passive acoustic localization (Dosso and Wilmut, 2011;Spiesberger, 2005;Tollefsen and Dosso, 2010). The essence of Bayesian modeling is to (i) express (known) measured variables as functions of (unknown) latent variables, (ii) assign a prior distribution to the latent variables, (iii) calculate a mathematical expression of the posterior distribution of the latent variables, and (iv) use numerical methods to compute posterior estimates of the latent variables.…”