Skip to main content Skip to main navigation

Publikation

Leveraging Artificial Intelligence for Optimal RAN Functional Split Decisions in 6G Space-Air-Ground Integrated Networks

Shama Noreen; Hans Dieter Schotten
In: 2025 IEEE Future Networks World Forum. IEEE Future Networks World Forum (FNWF-2025), November 10-12, Bangalore, India, IEEE, 2025.

Zusammenfassung

As non-terrestrial platforms expand 6G coverage, dynamically orchestrating radio access network (RAN) functional splits (FS) within disaggregated RAN architectures becomes critical for meeting stringent latency and reliability requirements. Conventional optimization methods fall short in addressing the uncertainty and heterogeneity of space–air–ground integrated networks (SAGIN), making artificial intelligence (AI) particularly well-suited to this orchestration problem. This paper provides a comprehensive roadmap for AI-driven decision-making of RAN FS in 6G SAGIN. It identifies architectural enablers specific to FS in SAGIN, formalizes key design objectives, and introduces a novel state-feature taxonomy. Leveraging federated learning (FL), it introduces a distributed, open RAN (O-RAN) compatible framework that supports real-time and resilient orchestration across satellites, aerial, and terrestrial nodes, while addressing key challenges in client scheduling, resource optimization, and the mitigation of stale updates.