Citation

Wu J, Xu H, Zheng Y, Tian Z. Accid. Anal. Prev. 2018; 121: 238-249.

Affiliation

University of Nevada, Reno 1664 N. Virginia Street, MS258, Reno, Nevada, 89557, United States. Electronic address: zongt@unr.edu.

Copyright

(Copyright © 2018, Elsevier Publishing)

DOI

10.1016/j.aap.2018.09.001

PMID

30265910

Abstract

Safety evaluation based on historical crashes usually has a lot of limitations. In previous studies, near-crashes are considered as surrogate data for safety evaluation. One challenge for the use of near-crashes data is the difficulty of data collection. The driving simulators and naturalistic driving data may not be suitable for safety evaluation at specific sites. The observational site-based methods such as human observers and video analysis also suffer from some limitations such as long time data processing or reduced performance influenced by weather or light condition. The roadside Light Detection and Ranging (LiDAR)-enhanced infrastructure provides a new solution for real-time data collection without the impact from weather or light. The high-resolution trajectories of all road users can be obtained from roadside LiDAR data. This paper aims to fill these gaps by presenting a method for near-crash identification based on the trajectories of road users extracted from roadside LiDAR data. This paper focused on vehicle-pedestrian near-crash identification particularly considering the increased risk of vehicle-pedestrian conflicts. Three parameters: Time Difference to the Point of Intersection (TDPI); Distance between Stop Position and Pedestrian (DSPP); Vehicle-pedestrian speed-distance profile, were developed for vehicle-pedestrian near-crash identification. The authors also recommended the thresholds for risk assessment of pedestrian safety. This method was coded into an automatic procedure for near-crash identification. This method is expected to significantly improve the current evaluation of pedestrian safety.

Copyright © 2018 Elsevier Ltd. All rights reserved.

Language: en

Keywords

Near crash; Pedestrian safety; Roadside LiDAR