Citation

Sengupta A, Ilgin Guler S, Gayah VV, Warchol S. Accid. Anal. Prev. 2024; 203: e107614.

Copyright

(Copyright © 2024, Elsevier Publishing)

DOI

10.1016/j.aap.2024.107614

PMID

38781631

Abstract

Vulnerable Road Users (VRUs), such as pedestrians and bicyclists, are at a higher risk of being involved in crashes with motor vehicles, and crashes involving VRUs also are more likely to result in severe injuries or fatalities. Signalized intersections are a major safety concern for VRUs due to their complex dynamics, emphasizing the need to understand how these road users interact with motor vehicles and deploy evidence-based safety countermeasures. Given the infrequency of VRU-related crashes, identifying conflicts between VRUs and motorized vehicles as surrogate safety indicators offers an alternative approach. Automatically detecting these conflicts using a video-based system is a crucial step in developing smart infrastructure to enhance VRU safety. However, further research is required to enhance its reliability and accuracy. Building upon a study conducted by the Pennsylvania Department of Transportation (PennDOT), which utilized a video-based event monitoring system to assess VRU and motor vehicle interactions at fifteen signalized intersections in Pennsylvania, this research aims to evaluate the reliability of automatically generated surrogates in predicting confirmed conflicts without human supervision, employing advanced data-driven models such as logistic regression and tree-based algorithms. The surrogate data used for this analysis includes automatically collectable variables such as vehicular and VRU speeds, movements, post-encroachment time, in addition to manually collected variables like signal states, lighting, and weather conditions. To address data scarcity challenges, synthetic data augmentation techniques are used to balance the dataset and enhance model robustness. The findings highlight the varying importance and impact of specific surrogates in predicting true conflicts, with some surrogates proving more informative than others. Additionally, the research examines the distinctions between significant variables in identifying bicycle and pedestrian conflicts. These findings can assist transportation agencies to collect the right types of data to help prioritize infrastructure investments, such as bike lanes and crosswalks, and evaluate their effectiveness.

Language: en

Keywords

Bicycle crash; Imbalanced data; Machine learning; Pedestrian crash; Surrogate measures of safety