Traffic Density Classification Based on V2X Communication Using XGBoost Algorithm

This study aims to develop a traffic density classification model based on V2X communication simulation using the Extreme Gradient Boosting (XGBoost) algorithm. A simulation-based quantitative approach was employed, utilizing Simulation of Urban Mobility (SUMO) to generate vehicle mobility data and Network Simulator 3 (NS-3) to simulate Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) communications. Mobility and communication data were integrated using the simulation and step attributes, then aggregated to obtain average Quality of Service (QoS) parameters at each timestep. The features included average delay, average packet loss, average throughput, and average jitter. Traffic density levels were determined using the quantile method and categorized into three classes: Low, Medium, and High. The results indicate that the V2I model outperformed the V2V model, achieving accuracies of 78.79% and 59.42%, respectively. Feature importance analysis identified average delay as the most influential feature in the classification process for both communication scenarios. These findings demonstrate that integrating SUMO, NS-3, and XGBoost can effectively classify traffic density based on V2X communication QoS parameters. This approach has the potential to support the development of intelligent transportation systems by using vehicle communication data as a basis for predictive solutions for traffic monitoring and management. Furthermore, it offers opportunities for adaptive