V2X Communication Reliability Improvement through QoS Prediction and Reservation Using Network Virtualization Technology

Due to the limited sensing range of vehicles, cooperative autonomous driving is being extensively researched to improve traffic safety and efficiency. This approach relies on sharing sensory data captured by surrounding vehicles and roadside units via wireless communication. By aggregating this data on a centralized server and delivering integrated environment models back to the vehicles, drivers and automated systems can perceive obstacles beyond the line-of-sight of their onboard sensors. However, if the number of vehicles within the server’s management area exceeds the available communication capacity, network congestion occurs, making it difficult to guarantee the Quality of Service (QoS). Furthermore, vehicles currently lack a mechanism to determine in advance whether their required QoS will be sustained. This study proposes a novel framework that leverages network virtualization technology to centrally manage the network via software, anticipate vehicle density, and predict QoS proactively. In addition, we introduce a QoS reservation mechanism to guarantee future network performance for individual vehicles. Network simulations demonstrate that the proposed method effectively prevents communication latency from degrading for vehicles with reserved QoS, even under heavily congested network conditions.