Design and Optimization of Sixth Generation (6G) Network Architectures Based on Distributed Artificial Intelligence and Federated Learning for Hyper-Global and Omnipresent Applications

Design and Optimization of Sixth Generation (6G) Network Architectures Based on Distributed Artificial Intelligence and Federated Learning for Hyper-Global and Omnipresent Applications

Alaedin Salahi1

1) MSc in Telecommunications Engineering, Sharif University of Technology, Tehran, Iran

IOI Code : AAHD-ACZZCK
Publication : National Conference on Information Technology, Nanotechnology, Artificial Intelligence and Technological Futures Studies(itnaf.ir)
Abstract :
    Considering the requirements of Sixth Generation (6G) networks to support hyper-global and omnipresent applications (such as pervasive metaverse, telesurgery, and planetary-scale industrial control), traditional centralized architectures face serious challenges in terms of latency, privacy, scalability, and energy consumption. This paper designs and optimizes a novel 6G network architecture based on Distributed AI and Federated Learning. The proposed framework, utilizing an Intelligent Edge Layer and a hierarchical structure of learning nodes, enables training of global models without transmitting raw user data. To optimize performance, an adaptive algorithm for resource allocation and coordination between nodes based on geographical dynamics and channel quality (QoS) is developed. Simulation results show that the proposed architecture achieves a 42% reduction in end-to-end latency, a 35% improvement in energy efficiency, and a high prediction accuracy of 96%, respectively, compared to centralized and distributed deep learning-based approaches, even in scenarios with severe data heterogeneity. Model convergence analysis also confirms the robustness of the proposed method in high-mobility networks. This research demonstrates that the combination of federated learning with distributed AI controllers provides a practical path to realize the vision of “ubiquitous and connected intelligence” in 6G communications.

Keywords : 6g networks distributed artificial intelligence federated learning edge computing resource allocation