|کد مقاله||کد نشریه||سال انتشار||مقاله انگلیسی||ترجمه فارسی||نسخه تمام متن|
|571908||877327||2016||7 صفحه PDF||ندارد||دانلود رایگان|
• HSM urban arterial crash prediction models lack the effect of horizontal alignment.
• Ratios of observed to predicted crashes for sites with/without curves were studied.
• Safety effect of horizontal curves was found significant on 2-lane urban arterials.
• For highways rather than 2-lane urban arterials further studies are suggested.
The crash prediction models of the Highway Safety Manual (HSM), 2010 estimate the expected number of crashes for different facility types. Models in Part C Chapter 12 of the first edition of the HSM include crash prediction models for divided and undivided urban arterials. Each of the HSM crash prediction models for highway segments is comprised of a “Safety Performance Function,” a function of AADT and segment length, plus, a series of “Crash Modification Factors” (CMFs). The SPF estimates the expected number of crashes for the site if the site features are of base condition. The effects of the other features of the site, if their values are different from base condition, are carried out through use of CMFs. The existing models for urban arterials do not have any CMF for horizontal curvature. The goal of this research is to investigate if the horizontal alignment has any significant effect on crashes on any of these types of facilities and if so, to develop a CMF for this feature.Washington State cross sectional data from the Highway Safety Information System (HSIS), 2014 was used in this research. Data from 2007 to 2009 was used to conduct the investigation. The 2010 data was used to validate the results. As the results showed, the horizontal curvature has significant safety effect on two-lane undivided urban arterials with speed limits of 35 mph and higher and using a CMF for horizontal curvature in the crash prediction model of this type of facility improves the prediction of crashes significantly, for both tangent and curve segments.
Journal: Accident Analysis & Prevention - Volume 95, Part A, October 2016, Pages 20–26