Automated End-To-End AI/ML Lifecycles for Radio Management in 6G Networks

Intelligent radio resource management is expected to play a vital role in addressing the strict user demands, possessing new challenges in the era of 6G services. In this paper, we provide an extensive study of artificial intelligence and machine learning (AI/ML) lifecycles, within the open radio access network (O-RAN) framework with particular emphasis on radio resource management advancement. More specifically, considering a multi-layered 6G network system, the AI/ML Cross-Layer Platform (AI-CLatform) is introduced to leverage AI/ML lifecycles devoted for O-RAN operation. With the near real-time RAN intelligent controller (Near-Rt RIC) devoted for radio intelligence, special emphasis is given on the development, deployment, optimization, and continuous monitoring of ML models within O-RAN, whereas the interactions between O-RAN and AI-CLatform are justified. To concretely illustrate the proposed end-to-end AI/ML sequential process, we present a proof-of-concept (PoC) practical scenario focusing on intelligent beamforming optimization for proactively managing the interference using recurrent neural networks (RNNs). The quantitative simulation findings prove the potential of the proposed AI/ML framework in enhancing critical 6G network functions within the O-RAN paradigm.