Estimating the Number and Locations of Boundaries in Reverberant Environments with Deep Learning

TL;DR

This paper introduces an improved deep learning approach for estimating the number and locations of reflective boundaries in reverberant underwater environments, eliminating the need for prior boundary information and retraining for different scenarios.

Contribution

The proposed method advances previous work by removing the requirement for prior boundary knowledge and retraining, enabling estimation of environments with one or two boundaries using deep learning.

Findings

Outperforms previous methods in simulation and real environments

Does not require prior boundary information or retraining

Capable of estimating environments with one or two boundaries

Abstract

Underwater acoustic environment estimation is a challenging but important task for remote sensing scenarios. Current estimation methods require high signal strength and a solution to the fragile echo labeling problem to be effective. In previous publications, we proposed a general deep learning-based method for two-dimensional environment estimation which outperformed the state-of-the-art, both in simulation and in real-life experimental settings. A limitation of this method was that some prior information had to be provided by the user on the number and locations of the reflective boundaries, and that its neural networks had to be re-trained accordingly for different environments. Utilizing more advanced neural network and time delay estimation techniques, the proposed improved method no longer requires prior knowledge the number of boundaries or their locations, and is able to estimate two-dimensional environments with one or two boundaries. Future work will extend the proposed method to more boundaries and larger-scale environments.