EUVIMEDEUVIMED entdecken
Crossref

A Lightweight AI-Powered Framework for Adaptive Multimodal Communication in Underwater Networks

Filippo Bragato, Filippo Donegà, Filippo Campagnaro, Michele Zorzi

IEEE Transactions on Communications · 2026 · 74 · 15995-16010

DOI: 10.1109/tcomm.2026.3742871 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Communications, 2026. Verfasst von Filippo Bragato und 3 weiteren Autor:innen. Die Arbeit nennt 34 Literaturangaben. Gefördert von: PRIN project BEASTIE, Italian Ministry of University and Research, European Union under the Italian National Recovery and Resilience Plan (NRRP) of NextGeneration EU, partnership on “Telecommunications of the Future”.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Underwater acoustic communication is an established transmission technology that enables numerous applications, including coastal surveillance, monitoring and maintenance of shore and offshore infrastructures, diver-to-diver communication, water quality assessment, and tsunami prediction, among many others. Although this technology allows long-range communication links (up to several kilometers), the challenges imposed by the acoustic channel—characterized by low bitrate and long propagation delay—make it difficult to support data-demanding applications such as, for example, real-time video transmission for piloting Work Class Remotely Operated Vehicles (ROVs). Conversely, alternative technologies such as optical, electromagnetic Radio Frequency (RF), and Magneto Inductive (MI) communication provide significantly higher rates at shorter distances. Among these, optical is considered as the most promising for underwater video transmission as it significantly outperforms RF in terms of range and MI in terms of bitrate. It is clear that no single underwater wireless communication technique is universally superior. Therefore, integrating multiple communication technologies within multimodal underwater networks offers the most effective approach to optimizing the performance. In these networks, it is especially important to understand how to combine transmission technologies optimally, and select which one to use depending on channel conditions and the desired Quality of Service (QoS) for the intended application. Traditional solutions select the best mode by either observing received power or probing channels periodically, whereas recent solutions focus on resource-demanding machine learning algorithms for channel selection. The main contribution of this work is to present the Multimodal Underwater Learning Framework (MULF), a lightweight QoS-based Reinforcement Learning (RL) algorithm that optimally selects the best transmission medium in multimodal underwater networks, implemented and evaluated as a C++ module for the open-source DESERT Underwater Framework. The results highlight how MULF outperforms traditional approaches while remaining suitable for onboard deployment on lightweight single-board computers used in many battery powered underwater bottom nodes and Autonomous Underwater Vehicles (AUVs), requiring no additional memory and increasing CPU usage by only 6% even on the highly resource-constrained Raspberry Pi Zero.

Förderung

  • PRIN project BEASTIE
  • Italian Ministry of University and Research
  • European Union under the Italian National Recovery and Resilience Plan (NRRP) of NextGeneration EU, partnership on “Telecommunications of the Future”
  • Italian National Plan for Military Research
  • Italian General Directorate of Naval Armaments
  • Italian Ministry of Defence

Zitieren

APA

Bragato, F., Donegà, F., Campagnaro, F., & Zorzi, M. (2026). A Lightweight AI-Powered Framework for Adaptive Multimodal Communication in Underwater Networks. IEEE Transactions on Communications, 74, 15995-16010. https://doi.org/10.1109/tcomm.2026.3742871

Vancouver

Bragato F, Donegà F, Campagnaro F, Zorzi M. A Lightweight AI-Powered Framework for Adaptive Multimodal Communication in Underwater Networks. IEEE Transactions on Communications. 2026;74:15995-16010. doi:10.1109/tcomm.2026.3742871

BibTeX
@article{Bragato2026,
  author = {Bragato, Filippo and Donegà, Filippo and Campagnaro, Filippo and Zorzi, Michele},
  title = {A Lightweight AI-Powered Framework for Adaptive Multimodal Communication in Underwater Networks},
  journal = {IEEE Transactions on Communications},
  year = {2026},
  volume = {74},
  pages = {15995--16010},
  doi = {10.1109/tcomm.2026.3742871}
}
Kostenlos starten →