An integrated clustering and Bayesian approach to investigate the severity of pedestrian collisions at highway-railway grade crossings collisions

Abstract

This study aims at developing a solid understanding of the contributing factors to pedestrian fatal and injury collisions at highway-railway grade crossings (HRGC), along with the impact of different warning devices that are commonly used at HRGCs. The study utilized integrated Machine Learning and Bayesian models to analyze the United States HRGC collision using the Federal Railroad Administration database between 2009 and 2018. The results demonstrate the association between different factors and the collision severity in each cluster and attempt to explain the inconsistency associated with the impact of some factors, such as weather conditions and pedestrian traits, on collision severity. The results also highlighted the conditions at which the different types of countermeasures and warning devices are most effective and the circumstances that limit their benefits. The results confirmed the benefits of the proposed analysis approach, in which collision data are classified into a group of clusters first before investigating the impact of the different factors on collision severity. The results wills support engineers and planners to develop specific policies and designs that aim at mitigating severe collisions at HRGCs and enhance pedestrian safety.

Keywords:

• severity of pedestrian collisions
• highway-railway grade crossings
• Bayesian techniques
• categorical principal component analysis
• latent class analysis
• Poisson-lognormal model

Disclosure statement

There are no relevant financial or non-financial competing interests to report.

Additional information

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.