Risk Driving Style Identification and Typical Pattern Characterization of Truck Drivers Based on Natural Driving Data

In recent years, China has experienced a sustained increase in the number of freight vehicles. However, due to their significant payload capacity and large physical dimensions, accident rates and fatalities remain high. The driving behavior of truck drivers is an important safety factor. The purpose of this study is to establish a method for evaluating driving styles based on driving behavior and express the micro driving behavior characteristics of drivers with different styles. All data in this study were obtained from natural driving data of truck drivers collected through GPS devices. Drivers were classified using a safety assessment system based on Data Threshold Variation (DTV) and Phase Plane Analysis with Limits (PPAL). Subsequently, data encoding was employed to construct a driving behavior characteristic graph based on micro driving behavior features. The results indicate significant variations in behavior graphs among drivers with different driving styles, and aggressive drivers tend to take multiple risky behaviors at the same time. This holds significant implications for distinguishing driving styles and facilitating fleet training and management.