Semantic Communication in Underwater IoT Networks for Meaning-Driven Connectivity

Ruhul Amin Khalil, Asiya Jehangir, Hanane Lamaazi, Sadaf Rubab, Nasir Saeed · 2026

Read Semantic Communication in Underwater IoT Networks for Meaning-Driven Connectivity by Ruhul Amin Khalil, Asiya Jehangir, Hanane Lamaazi, Sadaf Rubab, Nasir Saeed (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-13289.txt.

Introduction

Ruhul Amin Khalil and co-authors send sense, not raw bits, through sea nets to save scarce room. The Internet of Underwater Things links sensors for sea watch, reef health, and deep search. Sea links are harsh, though. Room is thin, lag is high, and power is tight. Raw data first nets waste all three.

Sense first links send what a read means, in brief form, in place of the full raw read. The team sets out how such links can work in sea nets. They weigh gains in room and power next to the risk of lost detail. For sea teams, the trade may be the price of nets that last.

Catalog record

CREATORRuhul Amin Khalil, Asiya Jehangir, Hanane Lamaazi, Sadaf Rubab, Nasir Saeed
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE59 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-13289

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Semantic Communication in Underwater IoT Networks for Meaning-Driven Connectivity Ruhul Amin Khalil, Asiya Jehangir, Hanane Lamaazi, Sadaf Rubab, Nasir Saeed Abstract The Internet of Underwater Things (IoUT) is revolutionizing marine sensing and environmental monitoring, as well as subaquatic exploration, which are enabled by interconnected and intelligent subsystems. Nevertheless, the underwater communication is bounded by narrow bandwidth, high latency, and strict energy constraints that are the source of efficiency problems on traditional data-centric networks. To tackle these problematic issues, this work provides a survey on the recent advancement of Semamtic Communication(SC) for IoUT—this is a novel communication paradigm that seeks to harness not symbol raw information but rather its meaning and/or contextual significance. In this paper, we investigate the emerging advanced AI-powered frameworks, including large language models (LLMs), diffusion-based generative encoders, and federated learning (FL), that bridge semantic compression with context-aware prioritization and robust information reconstruction over noisy underwater channels. Hybrid acoustic–optical–RF architectures and edge-intelligent semantic encoders are also considered as enablers of sustainable and adaptive operations. Examples in underwater archaeology, marine ecology, and autonomous underwater vehicles (AUVs) coordination are provided as a relief to illustrate the merits of meaning-driven connectivity. The paper concludes with some recommendations, including semantic representations standardization, cross-domain interpolation, and privacy-support schemes. These issues must be addressed in the future before trustworthy SC-enabled IoUT systems can be developed for underwater communication. I Introduction The IoUT is an emerging technology that interconnects underwater objects such as sensors, AUVs, and environmental monitoring systems to the Internet. This capability allows real-time data acquisition, remote observation, and smart decision-making across the marine research domain, the oil exploration & monitoring field, as well as environmental conservation activities. The most recent market analysis [ 1 ] has estimated that the world IoUT market was USD 1.22 million in 2024, and it is projected to grow at a compound annual growth rate of 16.15% between 2025 and 2034, with an estimated value of around USD 5.46 million by 2034 as depicted in Fig. 1 . This wide use reflects the increasing demand for timely and accurate information from the underwater space, for which conventional technologies largely fall short of expectations. From observing marine life and ecosystems to inspection of underwater installations (oil rigs, pipes), IoUT technologies are becoming vital for sustainable and safe operations [ 2 ] . Fig. 1 : Projected growth of IoUT market trends by 2034. Recently, the IoUT domain has encountered many practical implementation problems for communication. The underwater sensor nodes are usually powered by a battery, which can hardly be recharged [ 3 ] . A possible bottleneck to scalable and efficient communication is the low-power supply and large data transmission from IoT devices. For instance, sensors at depths may consume power between 10 mW and 10 W, making them hardly feasible to maintain continuous data transfer [ 4 , 5 ] . For addressing these challenges, intelligent communication strategies that can achieve the optimal energy consumption and avoid unnecessary data forwarding/redundant data delivery and paper-real-time communication are in great demand. Fig. 2 : Illustration of SC-based image transmission in an IoUT environment. One possible solution to most of these challenges is SC, which is a new communication paradigm that aims to transmit the semantics or value of the data rather than the raw data itself [ 6 ] . Traditional systems are centered around the precise transmission of bits and bytes, but this ultimately translates to inefficiencies, especially in resource-constrained environments like underwater systems. On the other end, SC wants to transmit pertinent information, thus enhancing the optimality of bandwidth use and energy consumption [ 7 ] . SC reduces the data transfer by sending only significant events, e.g., environmental anomalies or sensor threshold exceeded. In contrast, traditional methods mainly aim to reproduce accurate signals, while SC provides task-specific semantics and reduces redundant data by 80-99%. One of the examples is SC frameworks for underwater image transmission, in which vision-language models (VLMs), compress images down to 0.8-3.3\% of their original size by localizing regions of interest (like coral bleaching areas or pipeline cracks) and zeroing out irrelevant background regions [ 8 ] . This is consistent with the low data rate capabilities of acoustic modems, which face challenges in sending uncompressed high-resolution images ( 10 MB/im) in real time. Second, SC increases resilience to channel degradation: knowledge graph-informed decoders reconstruct textual emergency alerts with 92\% semantic similarity, despite significant reductions in signal-to-noise ratios (SNRs), as low as 5\mathrm{dB} , at which conventional error-correcting codes do not succeed below 10\mathrm{dB} . Some of the popular SC techniques and their corresponding performances are given in Table I . TABLE I : SC Technologies & Performance Metrics Technology Function Performance SAGE Framework [ 9 ] Image-to-text-to-image conversion 82% latency reduction vs. JPEG2000 AquaIntellect DNN-Bayesian [ 10 ] Optimal sensor placement 35% lower localization error Dual ControlNet [ 9 ] Detail-preserving image reconstruction 0.78 SSIM at 5 dB SNR CRITIC Methodology [ 11 ] Multi-attribute QoS optimization 0.92 decision satisfaction score In the IoUT, SC can entail a limitation in the transmission of continuous environmental measures (e.g., temperature or salinity). Instead, only important facts, such as a sustainable water temperature change or an abnormal salinity level, would be sent in a message. This is in terms of energy efficiency and bandwidth utilization, making the communication system more robust [ 12 ] . Fig. 2 shows a high-level view of an IoUT SC enabled image delivery. First, the deep semantic information of the underwater image is obtained through multi-scale feature extraction and semantic segmentation. These semantic properties are then delivered in the underwater acoustic channel, which can greatly reduce bandwidth demand [ 13 ] . At the receiver side, a semantic reconstruction network employs the received features to reconstruct the original image accurately. This strategy allows for efficient, reliable, and context-aware image transmission even under very harsh underwater conditions. Further, considering the context-aware SC into IoUT systems makes SC more adaptive due to watermark techniques that become variable according to the water depth, temperature, and condition of the network, either through error-correcting code mechanisms or by using “reference” in system design scenarios [ 9 ] . This flexibility is of special interest in applications with strict demands on real-time performance, e.g., AUVs, where mission-critical decisions need to be made from received sensor data. AI, ML, and SC could collaborate, which will result in more autonomous and reliable IoUT systems. This fusion not only improves the underwater communication performance but also brings us new opportunities for ocean study, environmental protection, and resource management. Table II provides a detailed comparison of traditional versus SC for IoUT network connectivity. TABLE II : Technical Comparison of Traditional communication vs. SC for IoUT networks. Parameter Traditional Communication Semantic Communication Improvement Factor Data Volume Transmits raw data (100% size) [ 9 ] Compresses to 0.8-3.3% of original size via semantic extraction [ 9 , 14 ] 30–125 \times reduction Transmission Efficiency 1-50 kbps acoustic links [ 15 , 14 ] Equivalent to 1.2-4 Gbps effective throughput through semantic compression [ 9 , 14 ] 24–80 \times bandwidth efficiency gain Robustness (5 dB SNR) Packet delivery ratio < 80% at 5 dB SNR [ 15 ] Maintains 0.85+ semantic similarity

Read the complete text