Predicting reduced visibility related crashes on freeways using real-time traffic flow data

    Authors

    H. M. Hassan;M. A. Abdel-Aty

    Comments

    Authors: contact us about adding a copy of your work at STARS@ucf.edu

    Abbreviated Journal Title

    J. Saf. Res.

    Keywords

    Real-time crash prediction; Freeways; Reduced visibility; Random; Forests; Matched case-control logistic regression; FORESTS; Ergonomics; Public, Environmental & Occupational Health; Social; Sciences, Interdisciplinary; Transportation

    Abstract

    Objectives: The main objective of this paper is to investigate whether real-time traffic flow data, collected from loop detectors and radar sensors on freeways, can be used to predict crashes occurring at reduced visibility conditions. In addition, it examines the difference between significant factors associated with reduced visibility related crashes to those factors correlated with crashes occurring at clear visibility conditions. Method: Random Forests and matched case-control logistic regression models were estimated. Results: The findings indicated that real-time traffic variables can be used to predict visibility related crashes on freeways. The results showed that about 69% of reduced visibility related crashes were correctly identified. The results also indicated that traffic flow variables leading to visibility related crashes are slightly different from those variables leading to clear visibility crashes. Impact on Industry: Using time slices 5-15 minutes before crashes might provide an opportunity for the appropriate traffic management centers for a proactive intervention to reduce crash risk in real-time. (C) 2013 National Safety Council and Elsevier Ltd. All rights reserved.

    Journal Title

    Journal of Safety Research

    Volume

    45

    Publication Date

    1-1-2013

    Document Type

    Article

    Language

    English

    First Page

    29

    Last Page

    36

    WOS Identifier

    WOS:000320429600004

    ISSN

    0022-4375

    Share

    COinS