Citation

Wu Y, Abdel-Aty MA, Park J, Selby RM. Transp. Res. C Emerg. Technol. 2018; 86: 97-110.

Copyright

(Copyright © 2018, Elsevier Publishing)

DOI

10.1016/j.trc.2017.10.025

PMID

unavailable

Abstract

Fog warning systems can convey warning messages to drivers and help to reduce crashes that may occur due to the sudden occurrence of low visibility conditions. This study aims to assess the effectiveness of real-time fog warning systems by quantifying and characterizing drivers' speed adjustments under different roadway types, traffic conditions, and fog levels. In order to explore how a driver perceives the fog warning systems (i.e., beacon and dynamic message signs (DMS)) when approaching a fog area, this paper divides the roads into three zones (i.e., clear zone, transition zone, fog zone) according to visibility levels and suggests a hierarchical assessment concept to explore the driver's speed adjustment maneuvers. For the three different zones, different indexes are computed corresponding to drivers' speed adjustments. Two linear regression models with random effects and one hurdle beta regression model are estimated for the indexes. In addition, the three models were modified by allowing the parameters to vary across the participants to account for the unobserved heterogeneity. To validate the proposed analysis framework, an empirical driving simulator study was conducted based on two real-world roads in a fog prone area in Florida. The results revealed that the proposed modeling framework is able to reflect drivers' speed adjustment in risk perception and acceleration/deceleration maneuvering when receiving real-time warning massages. The results suggested that installing a beacon could be beneficial to speed reduction before entering the fog area. Meanwhile, DMS may affect drivers' brake reaction at the beginning section of reduced visibility. However, no effects of warning systems for drivers' final speed choice in the fog can be observed. It is suggested that proper warning systems should be considered for different conditions since they have different effects. It is expected that more efficient technology can be developed to enhance traffic safety under fog conditions with a better understanding of the drivers' speed adjustments revealed in this study.

Language: en

Keywords

Hurdle beta regression model; Random parameter; Real-time fog warning system; Reduced visibility conditions; Speed adjustment