Integrating connected vehicle hard braking event data for proactive road network safety screening.

Abstract

Road network screening is traditionally based on analyzing historical crash data to identify road sections (hotspots) that could benefit from safety improvement. However, crash data analysis for road safety improvement is inherently reactive, given that crash risks are estimated in hindsight. This study examines the potential integration of hard braking events (HBE) from connected vehicles (CV) as crash surrogates for network screening. The common hotspots analysis showed that a significant proportion of the crash-based hotspots were also identified using HBE as follows: all crashes (86.25%), rear-end crashes (96.83%), no-injury crashes (85.37%), and severe injury crashes (78.22%). The comparison of the common hotspots ranking based on crash data and HBE revealed that the rear-end crashes had the least Total Rank Difference (TRD) for the top 20 and 30 hotspots. The average ranking difference shows a decreasing trend as the number of top hotspots increases for all the crash types except for severe crashes. This study provides helpful insight into the potential of HBE as crash surrogates for network screening. The findings can assist road safety management agencies in conducting timely road network performance evaluations and identifying locations that could benefit from advanced safety improvement to mitigate the possible manifestation of traffic conflicts into crashes.