Examining Left-Turn Speeds for Intersection Signal Timing Design Using Crowdsourced Trajectory Data

Abstract

Left-turn vehicles’ intersection entry speeds are critical in determining yellow and red clearance intervals. The signal timing design guidelines recommend collecting the left-turn entry speed or estimating it based on the speed limit. Estimations based on speed limits have generally been effective but can result in over- or underestimations at intersections with varying geometric characteristics. It is critical for agencies to identify geometric characteristics that could influence left-turning vehicles’ entry speeds. This awareness could inform decision making by pinpointing intersections requiring field data collection to estimate yellow and clearance intervals instead of estimation based on the speed limit. This paper aims to evaluate geometric characteristics influencing left-turning vehicle speeds. This study examined left-turning speeds at 60 signalized intersections in Tucson, Arizona. Crowdsourced trajectory data offer a cost-effective and scalable approach to capturing real-world driving behavior, making them particularly valuable for large-scale safety and operations studies. Ordinary least squares (OLS) and quantile regression models were used to analyze speed profiles and the influence of various factors. Findings revealed that the number of left-turn lanes, medians, and the width of the left-turn lane affect entry, mid-maneuver, and exit speeds for movements impeded by traffic queues at intersections. These results emphasize the importance of considering intersection geometry and speed limit during signal timing design. Future research could explore the impact of other vehicles, signal timing phases, demographic characteristics, and vehicle classifications to deepen our understanding of driver behavior in left-turn movements, thereby improving intersection design and traffic management strategies.