Hydra-RAN: Multi-Functional Communications and Sensing Networks for Collaborative-Based User Status
Rafid I. Abd, Kwang Soon Kim, Minsoo Kang · 2024
Conventional RANs typically adopt a one-size-fits-all approach, which limits their effectiveness in responding to the varying conditions inherent in dynamic modern environments. The rigid nature of these systems necessitates extensive manual tuning and configuration, which is both time-consuming and expensive. Therefore, there is an urgent need to develop an innovative network that combines diverse networks, services, and modern technology into a cohesive infrastructure. This convergence is essential for providing multi-faceted cooperation, cohesive functionality, and meeting contemporary applications in dynamic environments. The Hydra radio access network (H-RAN) has been conceptualized as a comprehensive platform. This innovative design aims to integrate all existing networks and services, establishing a cohesive environment where they can operate concurrently. H-RAN seeks to break down silos by fostering collaboration among diverse networks, facilitating seamless interoperability. This paper introduces a novel paradigm of H-RAN multi-faceted cooperation architecture that incorporates a dense deployment of sensor and radio units (SRUs) that work collaboratively to optimize user status decisions. In addition, we introduce inter-element cooperation in a cooperative multi-sparse input/multi-task learning-based federated learning paradigm, known as (C-SMTL), which is an integral component of the AI/ML D-engine allowing H-RAN components to align their objectives toward common goals, thereby optimizing learning outcomes. This collaborative focus paves the way for a more robust and efficient machine learning (ML) framework. A key highlight of the simulation findings is the approach's ability to increase classification accuracy by an impressive 95% while maintaining reliability.