Gap-acceptance Behavior and Safety Analysis in Roundabouts

Roundabouts have become essential tools for traffic engineers to improve safety and accommodate varying flow at intersections. However, due to the lack of empirical data, little effort has been made to analyze roundabouts from a combined perspective of traffic flow and safety. This study investigates the traffic conditions, gap acceptance behavior, and road safety at three Swiss roundabouts using drone-collected trajectory data. Moreover, we present a methodology to transform classical axis-aligned vehicle bounding boxes into rotated bounding boxes (RBB), representing vehicles more accurately. The fundamental diagram is used to identify traffic conditions, while the maximum likelihood method (MLM) is used for the estimation of the critical gap. The results suggest that traffic conditions significantly influence gap acceptance behavior, with drivers accepting shorter gaps as vehicle accumulation increases. A safety analysis is performed through the use of surrogate safety measures (SSMs), namely the Time-To-Collision (TTC) and Post-Encroachment Time (PET) indicators. Spatial distribution diagrams of TTC and PET events highlight critical areas at roundabout entrances and within the roundabout. In addition, vehicle-to-vehicle TTC interactions are modeled using a second-order polynomial regression, where the curvature at the vertex provides insight into the development of the event. Finally, an extensive evaluation of the added value of RBBs reveals the differences in SSM calculations, particularly in scenarios involving heavy vehicles and low-speed conditions, providing more representative measurements of vehicle interactions.