Paper Title: Federated Learning-Based Optimization for RIS-Enhanced 6G-V2X Communications
Conference Name: ICTC 2024The 15th International Conference on ICT Convergence 16-18 October 2024, Ramada Plaza Hotel
Abstract: Abstract—To address the performance challenges of reconfigurable intelligent surfaces (RIS) in full-duplex (FD) 6G-vehicle to everything communications(V2X), we propose a novel analytical framework. This framework evaluates an FD base station (BS) communicating simultaneously with one uplink (UL) and one downlink (DL) mobile user/vehicle, each assisted by a distinct RIS. Departing from traditional optimization techniques, we introduce an innovative Federated Learning (FL)-based method for superior performance. We formulate an optimization problem to jointly optimize the phase-shift matrices at both RISs, aiming to maximize the achievable sum-rate under discrete phaseshift constraints. The non-convexity of this problem is managed using Karush-Kuhn-Tucker (KKT) conditions. Our approach demonstrates significant enhancements in sum-rate performance. Comprehensive comparisons show that the FL-RIS-V2X approach outperforms DRL-RIS-V2X, DDQN-RIS-V2X, and NORIS-V2X configurations. Additionally, we explore the trade-off between uplink and downlink rates in FD systems, emphasizing the need for meticulous power management. Empirical results reveal substantial improvements in signal strength and system robustness, underscoring the efficacy of the proposed FL-RISV2X approach in optimizing V2X communications.