Analysis and Modeling of Crossing Behavior at Urban Intersections in China

Abstract

Crossing behaviors at intersections are conventionally described through the gap acceptance theory; however, many gap-forcing behaviors exist at Chinese urban intersections. In this study, gap-forcing behaviors were analyzed and modeled from the perspective of conflict avoidance using a new modeling methodology called virtual field graph. After providing the definition of the preemptive status and the method to calculate the preemptive level, a new driving behavior model that differentiates from the conventional gap acceptance model was established. Analysis of collected field conflict data indicates that 82.2% of drivers’ decisions are normal decisions, while 17.8% are abnormal decisions. This suggests that there should be a physiological critical preemptive level while the drivers are making decisions. The driving population’s critical preemptive level was calibrated with the collected field conflict data through the logistic regression method and maximum likelihood approach.