Citation

Moraldi F, Torre FL, Ruhl S. J. Transp. Saf. Secur. 2020; 12(8): 977-996.

Copyright

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

DOI

10.1080/19439962.2019.1571546

PMID

unavailable

Abstract

In Germany and Europe in general, a relevant proportion of all road fatalities occur in the nonmotorway rural road network. Although the predictive performance of most accident prediction models is unknown, the Highway Safety Manual (HSM) formulation offers a consistent method for making reliable crash frequency predictions. These considerations led to the development of a model to be used for making crash frequency on German rural two-lane, two-way road segments following the HSM approach. The first step was the development of the base model by fitting a base safety performance function (SPF) to a group of 206 base sections, approximately 61 miles long. The full model was subsequently obtained by combining the base SPF with the HSM Crash Modification Factors (CMFs) and calibrating the model using the full data set (10,169 sections approximately 2,463 miles long). The goodness of fit (GOF) of the base and the full model was tested, showing a good fit for both. The GOF of the simple application of the original HSM predictive method was also tested, showing a very bad fit. This led to the conclusion that the research conducted constitutes a model that combines the consistency of the HSM approach with a good fit to the German reality.

Language: en

Keywords

accidents; crash data; HSM; road safety models; SPF