5G-V2X Topology Intelligence: A GNN Approach

NR-V2X networks are intricate systems with dynamic topology, varied communication modes, and diverse requirements. They require minimal latency for safety-critical applications and rapid data speeds for entertainment. The complexity of V2X networks challenges conventional and heuristic optimisation techniques for resource allocation. These techniques often do not account for the networks’ dynamic nature, leading to less optimal performance in rapidly changing conditions. This paper briefly examines how Graph Neural Networks (GNNs) enhance resource management in V2X systems through graph-based modelling of connected devices. In particular, GNNs are employed to identify network conditions and mobility scenarios that require a handover to maintain optimal connectivity. Preliminary experimental tests confirm the validity of the proposed approach.