Analysis of Drivers’ Stop-or-Run Behavior at Signalized Intersections with High-Resolution Traffic and Signal Event Data

Abstract

Intersection safety is a national priority, and a driver's stop-or-run decision at signalized intersections is an important factor that can lead to red light running (RLR) and cause intersection-related accidents. This paper investigated this important issue by using high-resolution traffic and signal event data collected from loop detectors. First, a simple method was developed to identify first-to-stop, yellow light running, and RLR cases by using information from stop bar detectors located right behind the stop line and advance detectors located several hundred feet upstream from the stop line. Traffic data collected from advance detectors (including occupancy time and time gap), signal information collected from the signal system (including used yellow time and time left to yellow start), information from three preceding vehicles, and information from vehicles on adjacent lanes were all applied to identify the factors that significantly affect drivers’ decision making. A binary logistical regression method was applied to analyze the significance of all these factors. From the investigation results, it was found that occupancy time, time gap, used yellow time, time left to yellow start, time gap between first two preceding vehicles, whether the nearest preceding vehicle runs through the intersection during yellow, and whether there is a vehicle passing through the intersection on the adjacent lane showed a significant effect on drivers’ decisions. A prediction model, which predicted whether a driver stopped or ran through the intersection, was also developed with the information collected by advance detectors. The testing experiment showed that the model's accuracy was as high as 87%.