EUVIMEDEUVIMED entdecken
Crossref

A broadband modeling method for range-independent underwater acoustic channels using physics-informed neural networks

Ziwei Huang iD, Liang An, Yang Ye, Xiaoyan Wang, Hongli Cao, Yuchong Du, Meng Zhang

The Journal of the Acoustical Society of America · 01.11.2024 · 156 (5) · 3523-3533

DOI: 10.1121/10.0034458 ↗

Kurz gefasst

Fachartikel in der Zeitschrift The Journal of the Acoustical Society of America, online erschienen am 21.11.2024. Verfasst von Ziwei Huang und 6 weiteren Autor:innen. Beteiligte Einrichtung: Key Laboratory of Underwater Acoustic Signal Processing of Ministry of Education, Southeast University, Nanjing, 210096, China. Die Arbeit wurde bisher 33-mal zitiert (Crossref) und nennt 52 Literaturangaben. Gefördert von: Fundamental Research Funds for the Central Universities, National College Students Innovation and Entrepreneurship Training Program.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate broadband modeling of underwater acoustic channels is vital for underwater acoustic detection, localization, and communication. Conventional modeling methodologies, based on methods such as the finite element method, finite difference method, and boundary element method, generally facilitate computation for only a single frequency at a time. However, in broadband modeling, this characteristic presents limitations, requiring multiple computations across frequencies, thereby leading to significant time challenges. To solve this problem, we propose a rapid broadband modeling approach using physics-informed neural networks. By integrating the modal equation of normal modes as a regularization term within the neural network's loss function, the method can achieve rapid broadband modeling of underwater acoustic channel with a sparse set of frequency sampling points. Operating in range-independent underwater environments with a liquid semi-infinite seabed, the method proficiently predicts the channel response across the frequency band from 100 to 300 Hz. Compared to the results obtained from KRAKEN, our method improves computational speed by a factor of 25 at a propagation distance of 20 km, while maintaining a mean absolute error of 0.15 dB for the acoustic channel response.

Autor:innen und Institutionen

  • Key Laboratory of Underwater Acoustic Signal Processing of Ministry of Education, Southeast University, Nanjing, 210096, China.

ORCID-Profile: Ziwei Huang

Förderung

  • Fundamental Research Funds for the Central Universities
  • National College Students Innovation and Entrepreneurship Training Program

Zitieren

APA

Huang, Z., An, L., Ye, Y., Wang, X., Cao, H., Du, Y., & Zhang, M. (2024). A broadband modeling method for range-independent underwater acoustic channels using physics-informed neural networks. The Journal of the Acoustical Society of America, 156(5), 3523-3533. https://doi.org/10.1121/10.0034458

Vancouver

Huang Z, An L, Ye Y, Wang X, Cao H, Du Y, et al. A broadband modeling method for range-independent underwater acoustic channels using physics-informed neural networks. The Journal of the Acoustical Society of America. 2024;156(5):3523-3533. doi:10.1121/10.0034458

BibTeX
@article{Huang2024,
  author = {Huang, Ziwei and An, Liang and Ye, Yang and Wang, Xiaoyan and Cao, Hongli and Du, Yuchong and Zhang, Meng},
  title = {A broadband modeling method for range-independent underwater acoustic channels using physics-informed neural networks},
  journal = {The Journal of the Acoustical Society of America},
  year = {2024},
  volume = {156},
  number = {5},
  pages = {3523--3533},
  doi = {10.1121/10.0034458}
}
Kostenlos starten →