Enhancing 5G V2X URLLC Broadcast/Multicast Services With FL-Based Wireless Resource Allocation

Abstract

This paper addresses the challenges of wireless resource allocation for 5G Ultra-reliable Low-latency Communication (URLLC) broadcast/multicast services in Vehicle-to-Everything (V2X) scenarios. It proposes three key algorithms: an iterative resource allocation approach that decomposes optimization into power and spectrum subproblems, a federated learning-based multicast resource allocation scheme that protects data privacy while enabling distributed training, and a cooperative multi-agent reinforcement learning solution that treats vehicles as intelligent nodes to jointly optimize system throughput, URLLC delivery rate, and multicast performance. Path loss models, mobility patterns, and interference scenarios are analyzed for both unicast and multicast transmissions. Simulation results demonstrate that the proposed algorithms achieve superior performance in terms of throughput, reliability, and latency compared to traditional and baseline approaches.