Model-based convolutional neural network approach to underwater source-range estimation
This paper is part of a special issue on machine learning in acoustics. A model-based convolutional neural network (CNN) approach is presented to test the viability of this method as an alternative to conventional matched-field processing (MFP) for underwater source-range estimation. The networks ar...
Main Authors: | , |
---|---|
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Acoustical Society of America
2024
|
Subjects: | |
Online Access: | https://hdl.handle.net/1721.1/154268 |
id |
ftmit:oai:dspace.mit.edu:1721.1/154268 |
---|---|
record_format |
openpolar |
spelling |
ftmit:oai:dspace.mit.edu:1721.1/154268 2024-05-19T07:38:13+00:00 Model-based convolutional neural network approach to underwater source-range estimation Chen, R. Schmidt, H. 2024-04-23T17:41:18Z application/pdf https://hdl.handle.net/1721.1/154268 en eng Acoustical Society of America 10.1121/10.0003329 The Journal of the Acoustical Society of America 0001-4966 1520-8524 https://hdl.handle.net/1721.1/154268 . Chen, H. Schmidt; Model-based convolutional neural network approach to underwater source-range estimation. J. Acoust. Soc. Am. 1 January 2021; 149 (1): 405–420. Creative Commons Attribution https://creativecommons.org/licenses/by/4.0/ Acoustical Society of America Article http://purl.org/eprint/type/JournalArticle 2024 ftmit 2024-04-30T23:30:25Z This paper is part of a special issue on machine learning in acoustics. A model-based convolutional neural network (CNN) approach is presented to test the viability of this method as an alternative to conventional matched-field processing (MFP) for underwater source-range estimation. The networks are trained with simulated data generated under a particular model of the environment. When tested with data simulated in environments that deviate slightly from the training environment, this approach shows improved prediction accuracy and lower mean-absolute-error (MAE) compared to MFP. The performance of this model-based approach also transfers to real data, as demonstrated separately with field data collected in the Beaufort Sea and off the coast of Southern California. For the former, the CNN predictions are consistent with expected source range while for the latter, the CNN estimates have lower MAE compared to MFP. Examination of the trained CNNs' intermediate outputs suggests that the approach is more constrained than MFP from outputting very inaccurate predictions when there is a slight environmental mismatch. This improvement appears to be at the expense of decreased certainty in the correct source range prediction when the environment is precisely modeled. Article in Journal/Newspaper Beaufort Sea DSpace@MIT (Massachusetts Institute of Technology) |
institution |
Open Polar |
collection |
DSpace@MIT (Massachusetts Institute of Technology) |
op_collection_id |
ftmit |
language |
English |
description |
This paper is part of a special issue on machine learning in acoustics. A model-based convolutional neural network (CNN) approach is presented to test the viability of this method as an alternative to conventional matched-field processing (MFP) for underwater source-range estimation. The networks are trained with simulated data generated under a particular model of the environment. When tested with data simulated in environments that deviate slightly from the training environment, this approach shows improved prediction accuracy and lower mean-absolute-error (MAE) compared to MFP. The performance of this model-based approach also transfers to real data, as demonstrated separately with field data collected in the Beaufort Sea and off the coast of Southern California. For the former, the CNN predictions are consistent with expected source range while for the latter, the CNN estimates have lower MAE compared to MFP. Examination of the trained CNNs' intermediate outputs suggests that the approach is more constrained than MFP from outputting very inaccurate predictions when there is a slight environmental mismatch. This improvement appears to be at the expense of decreased certainty in the correct source range prediction when the environment is precisely modeled. |
format |
Article in Journal/Newspaper |
author |
Chen, R. Schmidt, H. |
spellingShingle |
Chen, R. Schmidt, H. Model-based convolutional neural network approach to underwater source-range estimation |
author_facet |
Chen, R. Schmidt, H. |
author_sort |
Chen, R. |
title |
Model-based convolutional neural network approach to underwater source-range estimation |
title_short |
Model-based convolutional neural network approach to underwater source-range estimation |
title_full |
Model-based convolutional neural network approach to underwater source-range estimation |
title_fullStr |
Model-based convolutional neural network approach to underwater source-range estimation |
title_full_unstemmed |
Model-based convolutional neural network approach to underwater source-range estimation |
title_sort |
model-based convolutional neural network approach to underwater source-range estimation |
publisher |
Acoustical Society of America |
publishDate |
2024 |
url |
https://hdl.handle.net/1721.1/154268 |
genre |
Beaufort Sea |
genre_facet |
Beaufort Sea |
op_source |
Acoustical Society of America |
op_relation |
10.1121/10.0003329 The Journal of the Acoustical Society of America 0001-4966 1520-8524 https://hdl.handle.net/1721.1/154268 . Chen, H. Schmidt; Model-based convolutional neural network approach to underwater source-range estimation. J. Acoust. Soc. Am. 1 January 2021; 149 (1): 405–420. |
op_rights |
Creative Commons Attribution https://creativecommons.org/licenses/by/4.0/ |
_version_ |
1799477627598143488 |