Modeling Fuel Consumption of Heavy-duty Trucks Using Telematics Data

Abstract

Fuel is the largest cost component in Vehicle Operating Costs (VOC) and a significant contributor to greenhouse gas (GHG) emissions in the trucking sector. This study developed a real-time telematics-based fuel consumption model for Euro-3 and Euro-4 trucks operating on toll roads in Indonesia, focusing on 5-axle heavy-duty trucks. The model utilizes telematics data, including average speed, gross vehicle weight, and road gradient under free-flow conditions, a novel aspect of this research. Two modeling approaches were applied: Model 1 employed multiple linear regression with Box-Cox transformation, while Model 2 utilized Generalized Linear Models (GLM) with a Gamma distribution and log link. Model 1 performed better, explaining 85.8% of the variability in fuel consumption (adjusted R² = 0.858) with a deviance of 0.947, RMSE of 0.033, and AIC of −3.246.625. Conversely, Model 2 recorded a deviance of 8.827, RMSE of 0.296, and AIC of −2.483. The Wilcoxon Signed Ranks Test indicated no significant differences between predicted and observed fuel consumption for both truck types, with a Z value of −1.700 (p = 0.089) for Euro-4 and −0.038 (p = 0.970) for Euro-3, supporting the model′s reliability. Beyond optimizing fuel consumption, the model offers practical recommendations for truck operators considering conversion to Euro-4 and provides valuable insights for policymakers developing energy efficiency strategies in the transportation sector. Further research is recommended to expand the model′s application to non-toll routes and integrate machine learning for more complex patterns.