Analysis of the impact of highway following behavior based on NGSIM on CO2 emissions

As global urbanization accelerates, the transportation sector has become a major source of greenhouse gas emissions. Following behavior, a common driving scenario on urban roads, significantly impacts traffic flow stability and vehicle emissions. This study utilizes the NGSIM dataset to extract tailgating segments involving passenger vehicles on highways. The MOVES model is employed to calculate the instantaneous CO2 emissions of vehicles before and during tailgating. Concurrently, an XGBoost model combined with the SHAP algorithm analyzes key factors influencing CO2 emissions during tailgating. Results indicate that the average speed difference between vehicles, the average headway distance, and the maximum headway time contribute most significantly to emissions. Based on these findings, this study proposes eco-following strategies for highways: maintaining the following vehicle's speed close to the leading vehicle's speed, keeping the average headway distance above 20 meters, and limiting the maximum headway time to within 2.5 seconds. This research provides data support and strategic references for optimizing micro-level traffic behavior and controlling vehicle emissions.