Distributed AI-Based Communication Systems for Sustainable Internet of Underwater Things

Next-generation networks aim to extend the Internet of Everything (IoE) vision to terrestrial, aerial, and underwater domains. However, when considering acoustic underwater sensing and communications, as traditionally done in the marine scenarios, challenges such as limited bandwidth, high latency, and energy constraints need to be faced. To address these limitations, this paper proposes a distributed AI-based framework for the realization of the Internet of Underwater Things (IoUT), enabling smart sensor nodes to perform localized data processing and adaptive communication. Our system integrates intelligent computing devices with underwater sensing capabilities, allowing cameraequipped nodes to classify marine species using lightweight AI models deployed directly on the nodes. Instead of transmitting raw data, nodes share only classification results, drastically reducing data volume. Software-defined radio (SDR) systems dynamically adapt transmission parameters (e.g., transmission power, modulation, etc.) to minimize energy consumption and environmental impact on marine ecosystems. This decentralized methodology enhances scalability, reduces latency, and saves resources while enabling real-time decision-making in dynamic environments. Performance evaluations using GNU Radio simulations prove reliable data delivery, measured by bit error rate (BER), throughput, and latency, while prioritizing marine ecosystem preservation. Results highlight the system's ability to balance efficiency, sustainability, and ecological conservation in next-generation underwater networks in view of realizing the vision of an Internet of Underwater Things.