5G/6G and its Association with Distributed Artificial Intelligence (DAI) Framework

The rise of new-generation mobile networks, including 5G and the impending 6G, poses formidable technical hurdles in attaining the ambitious benchmarks set by the research and industry communities. These challenges encompass accommodating a multitude of devices on a single network, ensuring ultra-reliable low-latency communication, sustaining adaptability and dynamism, and delivering ample high-quality bandwidth. To tackle these complexities effectively, there is an escalating demand for a unified approach amalgamating network management and control, featuring autonomous and adaptable actions. The presented Distributed Artificial Intelligence (DAI) framework harnesses Belief Desire Intention (BDI) agents endowed with machine learning capabilities, denoted as BDIx agents, because it uses ML under believes. These agents are dispersed across mobile devices, forming a multi-agent system (MAS) that incorporates Fuzzy Logic and Back-Propagation Neural Networks for Reinforcement Learning at the agents' perceptual and cognitive tiers. A prime illustration of the DAI framework is demonstrated in the context of Device-to-Device (D2D) communication within 5G and beyond networks. D2D communication's decentralized nature, coupled with a multitude of user devices (User Equipment or UEs), presents an ideal platform to showcase the capabilities of the DAI framework. By integrating BDIx agents into D2D UEs, it becomes possible to circumvent the conventional Base Station (BS) and establish direct links among neighboring UEs. This approach promises enhancements in spectral and energy efficiency, data rates, throughput, latency, interference management, and fairness. Given the manifold challenges introduced by D2D communication in 5G and 6G networks, the DAI framework is anticipated to play a pivotal role in surmounting these obstacles and fostering innovations in Artificial Neural Networks and other facets of these dynamic mobile networks. In my presentation, I will showcase the ADROIT6G EU project, which utilizes the proposed framework to realize BDIx agents. ADROIT6G's primary goal is to establish innovative research principles to advance low Technology Readiness Level technologies in preparation for the future 6G network architectures. This project seeks to enhance the current servicebased structures of 5G mobile networks by developing and validating a forward-looking, cognitive 6G architecture. This will be achieved through a fully distributed paradigm driven by Artificial Intelligence, deploying functional elements as virtual functions in cloud-native environments spanning the far-edge, edge, and cloud domains, and involving multiple stakeholders. These advancements aim to deliver improved performance, greater control, increased transparency in digital service interactions, support for innovative applications, and societal acceptance, marking a significant step towards the evolution of next-generation 6G networks