Unraveling Uncertainty in Eco-Driving: Exploring Inter-Driver and Intra-Driver Relationships in Fuel Consumption

Abstract

Uncertainty exists in driving behavior both between drivers (inter-drivers) and within individual drivers (intra-drivers). The proportions and relationships of these variations, especially across different driving conditions, remain unclear. This study investigates uncertainties in real-world traffic scenarios by classifying their sources, quantifying their relationships, and developing an optimized model for expected driving states. Driver behavior instability was characterized by analyzing the standard deviation of output power. Inter-driver behavioral uncertainty was assessed by analyzing operational data from different drivers, while intra-driver behavioral uncertainty was represented through data from the same driver across various trip conditions. Furthermore, an individualized stepwise optimization eco-driving model for varying driving conditions (ED-CIVDC), incorporating physiological and psychological factors, was developed to provide comprehensive feedback on driver behavior. Findings reveal that, within the speed range of 20 to 100 km/h, the ratio of intra-driver to inter-driver energy consumption uncertainty ranges from 0.228 to 0.558 and increases with speed. Moreover, drivers classified under the same type may exhibit diverse driving behaviors. A normal-type driver traveling at speeds of 30-35 km/h displays 16.3% aggressive behavior and 26.6% cautious behavior. The ED-CIVDC model reduces energy consumption by 0.96% to 11.48% after five optimization iterations, though its effectiveness diminishes at higher speeds.

Keywords: Eco-driving, Uncertainty, individual driver, individual expected driving state