EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Low-Complexity Multi-Step Signal Prediction for Underwater Acoustic Communications: A Joint Reservoir Computing and Transfer Learning Approach

Senthan Prasanth, Adnan A. Cheema, Gökhan Seçinti, Dac-Binh Ha, Berk Canberk, Octavia A. Dobre, Trung Q. Duong

IEEE Open Journal of the Communications Society · 2026 · 7 · 4147-4164

DOI: 10.1109/ojcoms.2026.3678450 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Open Journal of the Communications Society, 2026. Verfasst von Senthan Prasanth und 6 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Canada Excellence Research Chair (CERC) Program, Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery, Canada First Research Excellence Fund Transforming Climate Action.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Underwater acoustic communication faces significant challenges due to multipath propagation, signal attenuation, and environmental variability. Accurate signal prediction supports adaptive transmission by forecasting upcoming samples and enabling efficient resource allocation. This work presents a complexity-reduced echo state network (ESN) framework for underwater acoustic signal prediction using real-world shallow-water at-sea experimental data. By reducing the reservoir to 200 neurons and optimizing hyperparameters, the model achieves strong performance with low computational complexity. Single-step prediction on the first hydrophone (H1) yields a test ${R}^{{2}}$ of 0.9978. Furthermore, multi-step forecasting on H1 is performed using a recursive prediction strategy and evaluated up to 6 steps ahead, achieving ${R}^{{2}} \text {=} {0.8232}$ at step 6. Using transfer learning, we extend the H1-trained model across 14 additional hydrophones in the vertical array by retraining only the readout layer, which reduces retraining time and computational complexity. Cross-dataset validation between different shallow-water environments further demonstrates robustness to changes in propagation conditions. Across the vertical array, single-step transfer learning achieves ${R}^{{2}} \ge ~{0.997}$ for most hydrophones, while cross-dataset transfer yields ${R}^{{2}}$ values typically in the range 0.93-0.97.

Förderung

  • Canada Excellence Research Chair (CERC) Program
  • Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery
  • Canada First Research Excellence Fund Transforming Climate Action
  • Canada Research Chairs
  • Vietnam National Foundation for Science and Technology Development
  • Scientific and Technological Research Council of Türkiye (TÜBITAK) 1515 Frontier Research and Development Laboratories Support Program for Türk Telekom 6G Rese…

Zitieren

APA

Prasanth, S., Cheema, A. A., Seçinti, G., Ha, D., Canberk, B., Dobre, O. A., & Duong, T. Q. (2026). Low-Complexity Multi-Step Signal Prediction for Underwater Acoustic Communications: A Joint Reservoir Computing and Transfer Learning Approach. IEEE Open Journal of the Communications Society, 7, 4147-4164. https://doi.org/10.1109/ojcoms.2026.3678450

Vancouver

Prasanth S, Cheema AA, Seçinti G, Ha D, Canberk B, Dobre OA, et al. Low-Complexity Multi-Step Signal Prediction for Underwater Acoustic Communications: A Joint Reservoir Computing and Transfer Learning Approach. IEEE Open Journal of the Communications Society. 2026;7:4147-4164. doi:10.1109/ojcoms.2026.3678450

BibTeX
@article{Prasanth2026,
  author = {Prasanth, Senthan and Cheema, Adnan A. and Seçinti, Gökhan and Ha, Dac-Binh and Canberk, Berk and Dobre, Octavia A. and Duong, Trung Q.},
  title = {Low-Complexity Multi-Step Signal Prediction for Underwater Acoustic Communications: A Joint Reservoir Computing and Transfer Learning Approach},
  journal = {IEEE Open Journal of the Communications Society},
  year = {2026},
  volume = {7},
  pages = {4147--4164},
  doi = {10.1109/ojcoms.2026.3678450}
}
Kostenlos starten →