Interactive risk analysis on crash injury severity at a mountainous freeway with tunnel groups in China

Abstract

Traffic safety of freeways has attracted major concerns, especially for a mountainous freeway affected by adverse terrain conditions, constrained roadway geometry and complicated driving environments. On the basis of a comprehensive dataset collected from a mountainous freeway with a length of 61km but gathering 12 tunnels, this study seeks to examining the interactive effect of mountainous freeway alignment, driving behaviors, vehicle characteristics and environmental factors on crash severity. A classification and regression tree (CART) model is employed as it can deal with high-order interactions between explanatory variables.Results show that the driving behavior is the most important determinant for injury severity of mountainous freeway crashes, followed by the crash time, grade, curve radius and vehicle type. These variables, interacted with the factors of season and crash location, may largely account for the likelihood of high risk events which may result in severe crashes. Events associated with a notably higher probability of severe crashes include coach drivers involved in improper lane changing and other improper actions, drivers involved in speeding during afternoon or evening, drivers involved in speeding along large curve and straight segment during morning, noon or night, and drivers involved in fatigue while passing along the downgrade. Safety interventions to prevent severe crashes at the mountainous freeway include hierarchical supervision in terms of hazardous driving events, enhanced enforcement for speeding and fatigue driving, deployment of advanced driving assistance systems for fatigue driving warning, and cumulative driving time monitoring for long-distance-travel freight vehicles.