Citation

So JJ, Lim IK, Kweon YJ. Transp. Res. Rec. 2015; 2513: 56-62.

Copyright

(Copyright © 2015, Transportation Research Board, National Research Council, National Academy of Sciences USA, Publisher SAGE Publishing)

DOI

10.3141/2513-07

PMID

unavailable

Abstract

This study explored potential use of a traffic conflict-based surrogate safety assessment as an alternative way of evaluating the safety performance of roads and identifying potential sites for safety improvement. The study compared two major safety assessment streams (a statistical modeling method and a traffic conflict-based method) in their evaluation of crash risk to investigate the performance of the traffic conflict-based method as an alternative to the crash-based method in identification of hot spots. The empirical Bayes (EB) method coupled with the safety performance function (SPF), called the EB-SPF method, was used as a benchmark, and the conventional crash frequency method was used as a comparison supplement. These two methods were viewed as the crash-based methods because they relied on crash data. Traffic conflicts were estimated through the microscopic traffic simulation model VISSIM. The safety evaluation was performed separately for 24 signalized intersections and 86 segments in the Tysons Corner, Virginia, area. The safety measures estimated by the three methods (i.e., EB-SPF, crash frequency, and traffic conflict) were compared through Pearson correlation analysis, and hot spot identification results were compared through the use of rank-based mean absolute error values.

RESULTS showed that the conflict-based method had a fairly high correlation and a coefficient of 0.71 with the EB-SPF method in the resulting outcomes and performed better than the crash frequency method in identifying hot spots. Therefore, the conflict-based method can serve as a viable option for safety performance evaluation and hot spot identification, especially when sufficient crash data are not obtainable.