Horizontal curves present serious safety concerns since curves have a higher fatality rate than tangent roads. However, studies of curve crashes have been limited, due to lack of comprehensive curve dataset. The recent development of a statewide curve dataset has created the opportunity to examine broad curve safety in Florida. This study analyzed curve crashes and explored statistical machine learning methods to develop safety performance functions for curves. The study examined risk factors including traffic volume, demographics, roadway and curve characteristics and spatial relation of curves to intersections. The results outline high-risk locations and countermeasures for safety improvements on curves.
Xingjing Xu | | University of Florida
Sivaramakrishnan Srinivasan | Associate Professor | University of Florida