Crash Injury Severity Analysis of Familiar and Unfamiliar Drivers on Horizontal Curves of Rural Two-Lane Roads Using GIS and Random Forest Algorithm

Articles in Press

Document Type : Original Article

Authors

1 Department of Civil Engineering (Road and Transportation), Faculty of Engineering, University of Guilan, Rasht, Iran

2 Department of Civil Engineering (Road & Transportation), Faculty of Engineering, University of Guilan, Rasht, Iran

3 Department of Civil Engineering(Road and Transportation), Faculty of Engineering, University of Guilan, Rasht, Iran

10.22034/tri.2024.413567.3181

Abstract

By providing a transition between two straight road segments, horizontal curves play a significant role in the interurban road network. According to statistics, these locations are characterized by high incident and accident rates. The aim of this study is to introduce a geospatial based method for extracting the geometric parameters of horizontal curves on rural two-lane roads using Geospatial Information Systems (GIS) and satellite imagery, as well as identifying accident-prone curves through the utilization of spatial clustering functions and and investigating factors influencing the injury severity of accidents on the horizontal curves of rural two-lane roads based on familiar and unfamiliar drivers, using the Random Forest algorithm. The implementation results of the proposed method on the studied corridor (Rasht - Anzali) indicated that the most significant contributing factors to severity of crashes on accident-prone horizontal curves, which were identified using the kernel density estimation method were collision type and road light condition with coefficients of 0.51 and 0.41 for accidents involving familiar drivers and the factors of daily traffic volume and collision type with coefficients of 0.53 and 0.51 for accidents involving unfamiliar drivers. Factors such as daily traffic volume and driver age for familiar drivers, road surface conditions and road light condition for unfamiliar drivers were recognized as additional influential factors on the severity of crashes of studied corridor's horizontal curves.

Keywords

Main Subjects

Journal of Transportation Research

Articles in Press, Accepted Manuscript
Available Online from 21 January 2024

Files

Share

How to cite

Statistics