Determining Risk Factors for Vulnerable Road User Crashes: Integrating GIS and Binary Logistic Regression in a Medium-sized City of a Developing Nation

Abstract

The population increase in urban areas leads to heightened vehicle usage, resulting in more interactions among vehicles, pedestrians, and bicyclists, thereby raising substantial road safety issues. Infrastructure must be appropriately constructed for both motorized and non-motorized vehicles to mitigate safety problems for all road users. This study identifies high-crash-prone road segments and critical risk factors for vulnerable road user (VRU) crashes, encompassing pedestrians, bicyclists, and motorcyclists. This is accomplished by creating heatmaps from black-spot research through Geographic Information Systems (GIS) and by determining risk factors with a binary logistic regression (BLR) model. Crash data from Visakhapatnam, India, for the years 2014-2016 and 2019-2021 reveals that more than 50% of fatal incidents involved VRUs, and 75.6% of identified blackspot road segments are situated along a 62-km stretch of National Highway traversing the city. Roadways and land-use factors were collected during road-safety audits along the designated segment. The BLR Model identifies risk factors such as segment length, crash time, season, land use, driver sight distance, and vehicle type. The severity of crashes increases by 17.6% per unit increase in segment length, whereas insufficient visibility increases it by 43.1%. This integrated approach directs targeted interventions to improve the safety of VRUs.