Citation

Guo Y, Liu P, Wu Y, Chen J. J. Transp. Saf. Secur. 2020; 12(3): 419-440.

Copyright

(Copyright © 2020, Southeastern Transportation Center, and Beijing Jiaotong University, Publisher Informa - Taylor and Francis Group)

DOI

10.1080/19439962.2018.1490368

PMID

unavailable

Abstract

The primary objective of the study is to evaluate the impacts of right-turn treatments on right-turn-on-red (RTOR) conflicts at signalized intersections. Data were collected at 20 signalized intersections in Kunming, China. Three thousand, seven hundred and forty-six RTOR conflicts from five types of right-turn treatments were identified for analysis. Traffic conflict rates were compared among different types of right-turn treatments. The results showed that type 4 right-turn treatment (with raised channelized island and acceleration lane on the cross-street) has the lowest conflict rate, followed by the type 2 right-turn treatment (with painted channelized island and acceleration lane on the cross-street). Traffic conflict models were developed to investigate factors related to the RTOR conflicts frequency using full Bayesian estimation. Three types of models were developed and compared, including the fixed-parameter, the random-effect, and the random-parameter conflict models. The results showed that the random-parameter model outperformed the fixed-parameter model and the random-effect model. Further results from the traffic conflict model showed that the conflicting traffic volume, right-turn treatment type, right-turn radius and yield control sign for right-turn movement significantly affect the RTOR conflicts frequency. The elasticity results showed that the conflict frequency can be reduced by 15.03%, 28.4%, 18.53%, and 23.37% by type 1, type 2, type 3, and type 4 right-turn treatments, respectively.

Language: en

Keywords

full Bayesian estimation; Right-turn treatments; signalized intersection; traffic conflict model