摘要: |
Historically, project-level decisions for the selection of highway features to promote safety were based on either engineering judgment or adherence to accepted national guidance. These tools have allowed highway designers to produce facilities that have demonstrated an improving safety record in recent decades. However, these tools do not allow for comparison of the safety performance of dissimilar facilities or roadway attributes. To address this gap, researchers have been working for decades to develop Crash Prediction Models (CPMs) that can estimate, and ideally predict the expected safety performance of a highway based on its geometric and traffic control features. The main focus of this research was to evaluate the use of CPMs for rural two-lane highways in Kansas. Both CPMs provided in the Highway Safety Manual (HSM) and ones developed specifically based on Kansas data were considered. Many useful insights and tools were developed through this research study that focused on nonintersection related crashes. The primary conclusions were that single statewide calibration factors were calculated and recommended for rural two-lane highway segments and 3- and 4-leg stopped controlled intersections. A calibration function was also developed for highway segments that can be used to better account for animal crashes, which account for a significant number of rural two-lane highway crashes. |