Novel Approach to Modeling Impact of Driving Behavior on Fuel Consumption Based on Large Volume Real-Traffic Data

Abstract

This study examines the impact of driver behavior on vehicle fuel consumption using large volume telematics data. The author of the study developed a model based on the analysis of 1Hz GNSS position data, categorizing driving styles according to variations in speed and acceleration and comparing them with actual fuel consumption. The model employs speed band segmentation; however, unlike previous models, driver behavior is evaluated using rating curves derived from the probabilistic distribution of measured real-traffic data. The model differentiates between steady movement, acceleration, and deceleration of varying intensities to more accurately reflect driving efficiency. A simplified 1–5 scale is used to present results clearly. A long-term test conducted on a homogeneous fleet indicates that aggressive driving leads to increased fuel consumption and higher operational costs. In test vehicles, a one-point difference in the rating corresponded to a fuel consumption increase of up to 30%. These findings have practical applications for fleet management, driver training, fuel consumption optimization, utilization, and transportation policies aimed at enhancing efficiency and safety. Additionally, this study underscores the necessity of standardized metrics for different fuel types and provides a foundation for further research focused on improving vehicle efficiency and promoting safer driving practices.