INTRODUCTION The gap acceptance theory was primarily used to study pedestrian crossing behaviors, in accordance to static gaps that are calculated in the light of the cross section of crosswalk.… Click to show full abstract
INTRODUCTION The gap acceptance theory was primarily used to study pedestrian crossing behaviors, in accordance to static gaps that are calculated in the light of the cross section of crosswalk. However, pedestrians will face a series of dynamic gaps (especially at any uncontrolled multi-lane crosswalk) when they decide to cross the street, thus, pedestrians' decisions are made based on the dynamic gaps of each lane. METHOD Pedestrians' crossing behaviors at uncontrolled multi-lane mid-block crosswalk were investigated in this study. The lane-based gap (LGAP) was defined and five mid-block crosswalks were selected for observation in Wuhan, China. Pedestrians' behaviors and the corresponding traffic statuses were videoed as collected data, whose statistical analysis indicates that most pedestrians choose the rolling gap crossing strategy, which is different from existing research. Moreover, a logistic regression model was established to evaluate various influencing parameters (such as gender, age, waiting time and traffic volume) on the pedestrians' crossing strategy, whose accuracy is not satisfying. Therefore, the pedestrian dynamic gap acceptance (PDGA) model was put forward to describe pedestrians' crossing behaviors at any multi-lane crosswalk based on detailed analysis of the pedestrians' decision procedure. RESULTS The corresponding results show that its accuracy may be up to 88.6% to well describe pedestrians' crossing behaviors. CONCLUSIONS The PDGA model is appropriate to analyze pedestrians' dynamic decision procedures at multi-lane mid-block crosswalks. PRACTICAL APPLICATION The findings of this study can be used for safety and performance evaluation of crosswalks at mid-block locations in developing countries like China and India.
               
Click one of the above tabs to view related content.