Section‐Based Crash Risk Analysis Integrating the Effect of Traffic States and Road Geometry

ABSTRACT Highway collisions are influenced by a variety of factors, including dynamic traffic conditions and road geometry. A comprehensive understanding of how these factors specifically affect crash risk is essential for enhancing traffic safety. While previous studies have examined the relationship between traffic conditions and collision risk, as well as the influence of road geometry, limited attention has been given to analyses that consider both dimensions simultaneously. The configuration of road sections plays a critical role in vehicle behaviour and, consequently, in collision risk. This study introduces a section‐based crash risk analysis framework to investigate the interplay between traffic states and crash likelihood, with a particular focus on merging and diverging areas. Traffic states were classified using upstream and downstream detector speeds. Specifically, we analyse the impact of speed differences between upstream and downstream traffic, along with the influence of ramp flow on collision risk across various geometric configurations. Crash risk was quantified using crash occurrence (CR) and the potential crash occurrence rate (PCR). The relationships between traffic states and crash risk were modelled using polynomial and segmented regression. The results reveal that diverging sections exhibit the highest collision risk, especially under conditions of pronounced speed disparity, regardless of whether traffic is free‐flowing or congested. Moreover, the findings indicate a sharp increase in crash risk when the ramp‐to‐mainline flow ratio exceeds a critical threshold. These insights underscore the necessity of targeted traffic management strategies and optimized road design to mitigate high‐risk scenarios. They also emphasize the importance of future research that integrates both geometric and dynamic traffic characteristics in modelling collision risk.