Analysis of Factors Associated with Traffic Injury Severity on Urban Roads in Different Lighting Conditions

Abstract

This research has studied the severity of vehicle-to-vehicle accidents in Rasht City in various lighting conditions. In the present study, the target variable is the severity of an accident, which is split into two classes of no injury and injury/death. The effects of independent variables, including environmental, human, and roadway-related factors, on the vehicle-tovehicle accident severity, have been investigated. The classification and regression tree (CART) method has been used for modeling purposes. The results showed that on daylight accidents if the not-at-fault driver is a motorcycle, the accident tends to be fatal or injury with the probability of 95.6%. At night with enough lighting, if the at-fault vehicle is a motorcycle and the not-at-fault vehicle is not a motorcycle, there is a probability of 72% of injury or death in case of inattention to the front. Also, if the accident reason is inattention to the front and the not-at-fault driver is route familiar, there is a 95.1% probability of injury and death. In addition, at night accidents without enough lighting, if the not-at-fault vehicle is a motorcycle and the accident does not occur in the summer, with the 100% probability, the accident leads to injury and death.