Citation

Wu Y, Abdel-Aty M, Zheng O, Cai Q, Yue L. Transp. Res. Rec. 2019; 2673(4): 47-58.

Copyright

(Copyright © 2019, Transportation Research Board, National Research Council, National Academy of Sciences USA, Publisher SAGE Publishing)

DOI

10.1177/0361198119840617

PMID

unavailable

Abstract

A common type of bike lane at intersections is between the through lane and the right lane. With such design, right-turning drivers need to cross the bike lane to merge into the right lane, which could cause conflicts with bicycles on the keyhole bike lane. This study aims to develop a warning system for drivers to avoid vehicle-bicycle crashes in the bike lane area under a connected vehicle environment. To propose a reasonable warning system, 118 right-turning vehicle trajectories were collected by an unmanned aerial vehicle. Drivers' right-turning behaviors are investigated based on the trajectory data. Then, a vehicle-bicycle crash warning algorithm is proposed to calculate the post-encroachment time (PET) under different situations. By comparing the threshold value and the PET value, potential vehicle-bicycle crash locations in the bike lane area could be identified. The proposed algorithm is designed to be displayed on front windshields with an augmented reality display. The results suggested that the proposed algorithm could provide high prediction accuracy. Moreover, vehicle speed, vehicle location, bicycle speed, and bicycle location were found to have significant impact on the locations of dangerous areas. It is expected that the proposed warning system could be used to identify the dangerous areas and deliver warning information for right-turning drivers when they are approaching an intersection. The warning system could help drivers be more prepared for the upcoming right-turning maneuver, and thus improve traffic safety for both drivers and cyclists at intersections.

Language: en