AI-Driven 6G RAN Design for Efficient, Resilient and Environmentally Responsible Networks
Abstract
DRIVING-6G aims to develop and validate an AI/ML-driven framework for joint dynamic optimization of real-time sensing, computation offloading, communication and frequency/resource allocation in 6G RANs, leveraging cross-layer coordination protocols and decoupling ML algorithms from underlying RAN functions. In particular, DRIVING-6G will deliver solutions for multi-objective cross-layer RAN functions optimisations in real-time by integrating AI-powered cognition and collaborative intelligence. At the PHY/MAC, joint optimisations will be driven by advanced AI/ML technologies to orchestrate signal processing across the transmitter and receiver chains, enhancing system sustainability and computational feasibility for support of high throughput and low latency. Integration of sensing techniques with PHY layer optimisations will be implemented towards a ML-driven optimisation framework in support of multi-user resource allocation and for trustworthy communication links and increased reliability. AI and sensing-driven RAN optimisations will be achieved via sharing the intelligence among sensors for enhanced sensing capability.
