A New Statistical Model for Waveguide Invariant-Based Range Estimation in Shallow Water

TL;DR

This paper introduces a statistical model for waveguide invariant-based range estimation in shallow water, enabling passive localization using ship noise without detailed environmental knowledge, validated with real measurements.

Contribution

It develops a likelihood-based statistical model for WI-based range estimation in shallow water, improving passive acoustic localization methods.

Findings

Effective range estimation demonstrated with real ship noise data.

The model performs well in high signal-to-noise ratio scenarios.

Passive localization is achievable without detailed environmental parameters.

Abstract

Navigation and source localization in the undersea environment are challenged by the absence of a ubiquitous positioning system. Passive acoustic ranging offers a valuable means of obtaining location information underwater. We present a range estimation method based on waveguide invariant (WI) theory, using ship noise recorded by a hydrophone as an acoustic source. The WI is a scalar parameter that describes the interference patterns in spectrograms caused by the interaction of acoustic wave modes propagating in a waveguide, such as shallow water. WI theory enables ranging using a single receiver without detailed knowledge of the environment. In this paper, underwater acoustic signals radiated by a moving large ship, which include broadband and tonal components, are employed for WI-based ranging in a range-independent shallow water environment. In particular, we develop a likelihood function for WI-based range estimation by introducing a statistical model for high signal-to-noise ratio scenarios. Here, the broadband component is assumed to dominate the background noise. The effectiveness of the proposed range estimation method is demonstrated using real acoustic measurements of a moving container ship recorded during the Seabed Characterization Experiment 2017 (SBCEX17).