Predicting transmission loss in underwater acoustics using convolutional recurrent autoencoder network

Mallik, W. , Jaiman, R. K. and Jelovica, J. (2022) Predicting transmission loss in underwater acoustics using convolutional recurrent autoencoder network. Journal of the Acoustical Society of America, 152(3), pp. 1627-1638. (doi: 10.1121/10.0013894) (PMID:36182300)

Abstract

Underwater noise transmission in the ocean environment is a complex physical phenomenon involving not only widely varying physical parameters and dynamical scales but also uncertainties in the ocean parameters. It is challenging to construct generalized physical models that can predict transmission loss in a broad range of situations. In this regard, we propose a convolutional recurrent autoencoder network (CRAN) architecture, which is a data-driven deep learning model for learning far-field acoustic propagation. Being data-driven, the CRAN model relies only on the quality of the data and is agnostic to how the data are obtained. The CRAN model can learn a reduced-dimensional representation of physical data and can predict the far-field acoustic signal transmission loss distribution in the ocean environment. We demonstrate the ability of the CRAN model to learn far-field transmission loss distribution in a two-dimensional ocean domain with depth-dependent sources. Results show that the CRAN can learn the essential physical elements of acoustic signal transmission loss generated due to geometric spreading, refraction, and reflection from the ocean surface and bottom. Such ability of the CRAN to learn complex ocean acoustics transmission has the potential for real-time far-field underwater noise prediction for marine vessel decision-making and online control.

• Item Type: Articles
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Mallik, Dr Wrik
• Authors: Mallik, W., Jaiman, R. K., and Jelovica, J.
• College/School: College of Science and Engineering > School of Engineering > Autonomous Systems and Connectivity
• Journal Name: Journal of the Acoustical Society of America
• Publisher: Acoustical Society of America
• ISSN: 0001-4966
• ISSN (Online): 1520-8524
• Published Online: 13 September 2022
• Copyright Holders: Copyright © 2022 Acoustical Society of America
• First Published: First published in Journal of the Acoustical Society of America 152(3): 1627-1638
• Publisher Policy: Reproduced in accordance with the publisher copyright policy