Re-Clustering for High-Performance Metaverse in Vehicular Networks

The metaverse has the potential to provide new technology and application opportunities for vehicle networks. Metaverse experiences require seamless and high-resolution data flows to enable timely and accurate interactions. However, vehicle networks face frequent topology changes, easily suffering from data losses and delay variations. This paper studies novel clustering algorithms to enable vehicle networks to reliably achieve high-throughput and low delay jitter performance. We propose two vehicle clustering algorithms: the cluster-quality-based re-cluster decision algorithm (CQRD), and the link-quality-based recluster decision algorithm (LQRD). The proposed algorithms dynamically re-cluster vehicles in response to changing network conditions, thereby maintaining link stability, minimising jitter and improving overall throughput. Our extensive simulation results show that our algorithms outperform popular baseline approaches in terms of throughput and delay jitter without compromising on performance stability. These results indicate the potential of leveraging cluster and link quality metrics to support the metaverse in vehicle communication environments.