A Comparative Analysis of Surrogate Safety Indicators for Road Hazard Assessment Using Smartphone GPS Data

Road Safety Assessment (RSA) can be conducted through two primary approaches: (۱) the reactive approach, which relies on historical crash data; and (۲) the proactive approach, which employs Surrogate Safety Indicators (SSIs). In the reactive approach, hazard potential is evaluated directly by measuring accident frequency, whereas in the proactive approach, SSIs must be validated against actual safety measures. However, few studies have systematically compared multiple SSIs to determine which indicators provide the most reliable correlation with actual crash occurrence, particularly in rural contexts where data collection is challenging. The purpose of this research is to compare ۳۹ SSIs within the proactive RSA framework to identify potential hazards on rural road segments. An observational study was conducted employing Pearson correlation analysis, using vehicle velocity data collected via GPS-enabled smartphones from ۱۰۰ vehicles across ۱۲۵ km of rural roads. The analysis divided the road network into uniform ۵-km segments and evaluated various crash severity levels, including fatal, injury, and total crashes. Among ۳۹ proposed indicators, the deceleration frequency indicator (DNM۱, counting all decelerations during braking without predefined thresholds) demonstrates the strongest correlation with crash data (r = ۰.۸۰۳ for property-damage crashes), establishing it as a highly robust proactive safety measure for rural road networks. Additionally, results revealed that traditional metrics like average vehicle velocity alone were poor predictors of safety compared to comprehensive deceleration-based measures, emphasizing the necessity of granular kinematic data. The findings enable transportation agencies to implement cost-effective, smartphone-based safety monitoring systems for real-time identification of high-risk rural road segments.

کلیدواژه ها:

Road safety ، Crash ، Surrogate safety indicator (SSI) ، Deceleration number of vehicles ، Smartphones ، Real-time data