Sequential source localization in shallow water by single-receiver probabilistic focalization

In a shallow water waveguide, due to reflection and refraction, acoustic signals travel along multiple propagation paths from the transmitter to the receiver. Each propagation path can potentially provide additional information for source localization. We introduce a sequential Bayesian estimation method that can localize a continuous acoustic source in the water column using only a single mobile receiver by exploiting multipath propagation. Our probabilistic focalization approach estimates the time-varying range and depth of a continuous broadband acoustic source in the presence of measurement-origin uncertainty. In particular, cepstrum processing is used to extract time-difference-of-arrival (TDOA) measurements from the raw acoustic data, and probabilistic data association is performed to match extracted TDOA measurements with modeled TDOAs related to modeled pairs of propagation paths. The performance of our approach is evaluated in simulation and with real acoustic data recorded in a variety of scenarios.