A data-driven approach to assess the impact of road infrastructure on platoon cohesion and CO2 emissions

and improve traffic efficiency and safety. While platooning has been widely studied and successfully applied on highways, its adoption in urban environments remains limited due to unique challenges. Features such as intersections, traffic signals, and variable traffic conditions disrupt platoon cohesion, leading to increased acceleration, fuel use, emissions, and safety risks. In this study, we propose a novel routing framework that integrates road feature analysis to optimize platoon cohesion and emissions. Using real-world data collected from a test involving platoons of three and six vehicles operating under different traffic conditions, we apply linear regression and machine learning models to quantify the influence of traffic-related features on platoon cohesion and CO2 emissions. Based on this analysis, we develop a routing framework that integrates road network information to optimize routes either for inter-vehicle distance or for emission reduction. To validate the approach, we conduct experiments using both real-world and simulated urban routes. Results show that routing based on cohesion optimization improves platoon cohesion by 12% compared to conventional shortest-path routing. Similarly, emissions-optimized routes lead to a 5% reduction in CO2 emissions. These findings demonstrate that incorporating road feature analysis into the routing process can enhance the effectiveness of platooning in urban environments. Overall, the proposed approach provides a practical solution for improving platoon performance in cities and supports the broader deployment of energy-efficient, environmentally friendly automated vehicle technologies in complex traffic networks.