Modeling Pedestrian Crossing Behavior and Safety at Signalized Intersections

Pedestrian noncompliance behavior is one of the most critical causes of pedestrian involved traffic crashes at intersections in India. Thus, the objectives of this study are to examine how various factors affect pedestrian crossing behavior and to propose models for pedestrian crossing behavior and level of safety at signalized intersections which will be useful to regulate pedestrian flow. The data were collected with video and a user perceptions survey at six selected signalized intersections in Mumbai, India. The differences between pedestrian crossing behavior with respect to personal characteristics, socioeconomic attributes, and existing crossing facilities were identified using Pearson correlation and odd ratio tests. Furthermore, the major reasons for noncompliance behavior were obtained by analysis of field data to prevent noncompliance behavior and enhance pedestrian safety. The results showed that a significant number of the pedestrians violated the traffic signal to save time and for convenience (46%). A binary logit model was developed to evaluate the impacts of contributing factors on pedestrian crossing behavior. Further, an ordered probability model was established to evaluate and estimate the pedestrian level of safety at signalized intersections. Two models were validated, and their statistical results show that the models predict the pedestrian crossing behavior and safety level more precisely. Developed models and study outcomes can help transport planners and designers understand pedestrian crossing behavior on crosswalks at signalized intersections and thus create a safer crossing environment for all pedestrians.