Factors associated with vulnerable road user crashes on freeways in North Carolina

Abstract: A vulnerable road user (VRU) is a nonmotorist, such as a pedestrian, bicyclist, or person using a personal conveyance, including highway workers on foot in a work zone. In recent years, VRUs have accounted for a growing share of U.S. roadway fatalities. For example, bicyclist fatalities increased by 13 percent from 976 in 2021 to 1,105 in 2022, while pedestrian fatalities in 2022 (7,470) reached the highest level since 1981. VRU safety assessment is a critical component of the Strategic Highway Safety Plan (SHSP) in North Carolina (NC). Over the past decade, NC has experienced a substantial number of VRU crashes, with 22,057 pedestrian crashes and 8,927 bicycle crashes between 2012 and 2021. More importantly, A significant number of fatalities and serious injuries involve VRUs on freeways in NC, although VRUs are generally prohibited from freeways. For example, 6% of VRU crash incidents occurred on the interstate highway in NC, while only 1% of road miles are interstate highways. However, while much research has focused on VRU fatalities on urban roads, little is known about VRU fatalities and serious injuries on freeways.

To fill in this gap, this paper will investigate the VRU crash data from 2010 to 2022. The data were obtained from the NC Pedestrian and Bicycle Crash Data Tool. This paper will first identify hotspots of VRU fatalities and serious injuries on freeways using Geographic Information System and spatial point pattern mining. Then, to evaluate the impact of various risk factors on VRU crashes, considering freeway versus non-freeway locations and crash severity, we will first conduct an exploratory comparative analysis across four categories: (1) all VRU crashes on freeways, (2) VRU fatalities and serious injuries on freeways, (3) all VRU crashes on non-freeways, and (4) VRU fatalities and serious injuries on non-freeways. This analysis will consider a range of risk factors, including built environment characteristics (e.g., urban setting and land use), demographic information (e.g., race, age, and poverty level), time, weather conditions, road characteristics (e.g., number of lanes and speed limit), etc. Moreover, this paper will employ regression models to statistically infer the impact of these risk factors on VRU crashes on freeways, predicting the probability of crashes or crash counts based on these factors. The findings of this paper will provide critical insights into geographical distribution of freeway-specific VRU crashes and the underlying risk factors.