Assessment of eco-driving strategies on carbon emissions for hybrid vehicles through portable emissions measurement systems

Eco-driving is considered a cost-effective way to reduce fuel consumption and carbon emissions. However, eco-driving strategies for hybrid electric vehicles (HEVs) are understudied. Therefore, this study analyzed extensive road test data to assess HEV carbon reduction under different driving behaviors and to identify optimal eco-driving conditions. Firstly, the portable emissions measurement system (PEMS) was used to characterize the real-world emissions from two vehicles, one conventional vehicle (CV) and the other HEV. The results indicate that HEVs reduce average CO₂ emissions by 24.5%–54.7% compared to CVs. Secondly, based on the measured data, the impact of driving behavior on emission was investigated. It demonstrated that driving behavior was closely linked to engine operating state in HEVs, which in turn significantly affects carbon emissions. Notably, the emission reduction advantage of HEVs diminishes when considering only the engine-on state. At cruising speeds below 10 m/s, HEVs emit approximately 68% more CO₂ than CVs due to frequent start-stop cycles occurring. Finally, an eXtreme Gradient Boosting (XGBoost) model was proposed to predict engine operating status based on driving behavior and external traffic conditions. Combined with the Local Interpretable Model-Agnostic Explanation (LIME) algorithm, this model provides insights into the factors influencing engine state predictions, thus offering real-time eco-driving strategies for HEVs’ drivers. These findings reveal the carbon emission characteristics of HEVs under microscopic driving behavior and enhance the carbon reduction potential of HEVs in combination with eco-driving.