Pedestrian Injury Severity Analysis at Signalized Intersections: A Survival Analysis Based Approach

摘要

In order to investigate pedestrian injury severity at signalized intersections, a survival analysis-based approach is proposed to analyze injury severity varying with time to identify the influencing factors, while addressing the heterogeneity of unobserved factors at different signalized intersections. The crash data of Las Vegas metropolitan area from 2004 to 2008 is applied, involving 450 signalized intersections with 550 pedestrian crashes. To address the heterogeneity issue due to unobserved factors at different signalized intersections, the Weibull model with gamma heterogeneity is employed and compared with the Weibull model. The comparison indicates that heterogeneity is present in the Weibull survival analysis. The results show that pedestrian crashes, time of day, light conditions, and annual average daily traffic (AADT) on minor arterials are potentially significant factors of increasing the pedestrian injury severity probability. The findings provide useful insights for practitioners and policy makers to improve pedestrian safety at signalized intersections.