Assessing right-turning vehicle-pedestrian conflicts at intersections using an integrated microscopic simulation model

Abstract

Frequent vehicle-pedestrian conflicts deserve special attention for safety assessment at intersections. This study helps verify how the simulation as an innovative approach can be utilized for right-turning vehicle-pedestrian conflict assessment at intersection crosswalks prior to implementation. Various behavior models such as vehicle turning path, turning speed, gap acceptance model and pedestrian behavior model, have been established. Through integrating the calibrated models into one simulation platform, the stochastic behavior of vehicles and pedestrians under different geometric layouts and operational conditions can be reproduced. Based on the field data collected by an unmanned aerial vehicle (UAV) at two urban intersections in Beijing, China, it was demonstrated through validation of surrogate safety measures (SSMs), i.e., Post Encroachment Time (PET) and vehicle passing speed at conflict points, that the simulation model can reasonably represent the frequency and severity of conflict occurrence at signalized crosswalks. The sensitivity analysis results indicated that large dimensions and turning angles of intersections tend to result in undesirable safety performance.