Spatial Analyses of Traffic Conflicts to Assess Safety at Signalised Intersections

Abstract

Safety improvements at intersections play an important role in reducing the number and severity of crashes in roadway networks, as such locations are associated with high numbers of crashes. This paper presents an approach to assessing traffic safety at intersections by using GIS software tools to achieve spatial analyses of traffic conflicts data to enable the identification and ranking of intersections by risk instead of using crash data. Traffic conflict data were collected for nine four-legged signalized intersections in Baghdad city, Iraq through site-based observations in 2017, and archived crash data for three years obtained to conduct a comparison. Kernel density estimation was used as a spatial analysis tool in ArcGIS software to estimate conflict intensity at each location, and an Intersection Prediction Conflict Index (IPCI) used to represent the ratio of the proportion of conflicts events occurring within the identified hazard locations (hotspots within intersections) to the area covered. The results showed that the developed approach, based on the spatial properties of traffic conflict events, is efficient and reliable in terms of identifying potential hot spots and ranking sites as compared with observed crash frequency. This approach can thus be used as a tool to support effective decision making about improvements in safety at signalised intersections.