Risk-field assessment for signalized intersections using surrogate safety measure

Abstract

Introduction: Traffic signals are the controlling devices aimed to reduce crossing conflicts at intersections. However, rear-end and lane-changing conflicts at signalized intersection approaches are a significant problem. This work aims to proactively assess and spatially map the safety and risk at signalized intersection approaches by field data collection and microsimulation modeling using PTV-VISSIM^TM. Method: Three signalized intersections operating under mixed traffic conditions were selected and traffic data were collected from the field using video cameras. Post Encroachment Time (PET) was selected as the surrogate safety measure for data analysis. Critical PET values were obtained for each vehicle category by drawing cumulative frequency distribution plots. Microsimulation of the intersections was done using PTV-VISSIM^TM and the simulated trajectory data were exported to Surrogate Safety Assessment Model (SSAM). Risk-field models were then developed for different signal change intervals based on the number of critical conflicts. Results: Risk was found to be minimum near the stop-line. From the perspective of a driver approaching signalized intersection, as distance to stop-line reduces, risk increases up to a maximum value and then decreases as the driver gets closer to the stop-line. The location at which maximum risk occurs varies with intersections depending on its geometric and traffic characteristics. High risk area for all intersections was observed to be on middle lane, while considering the entire signal cycle and while considering green time alone. The high-risk area is found to shift toward the median-side lane while considering red time alone. While evaluating safety, it was observed that high-risk area shifts away from the intersection during green signal and shifts toward the intersection during red signal, compared to the risk during the entire signal cycle.