Multimodal Collaborative Perception for 6G V2X Power and Spectrum Management

The advent of sixth-generation (6 G) networks introduces significant advancements for Vehicle-to-Everything (V2X) communications that offer ultra-low latency, massive connectivity, and enhanced spectral efficiency. However, the dynamic and heterogeneous nature of V2X environments, coupled with the use of high-frequency bands like millimeter-wave (mmWave) and terahertz (THz), poses significant challenges for Radio Resource Management (RRM). Traditional RRM techniques frequently fail to adapt to rapidly changing network conditions, which results in suboptimal performance. Attending to these concerns, in this paper, we propose a novel AI-driven RRM framework for 6G V2X networks that integrates multimodal collaborative perception through LiDAR-based environmental information with traditional wireless metrics. The proposed framework aims to optimize data rate, energy efficiency, and fairness across the network. The results of our extensive simulations demonstrate that our approach significantly outperforms traditional RRM methods, confirming the effectiveness of incorporating multimodal collaborative perception data and advanced deep learning time series techniques.