In contrast to conventional methods, this study took a holistic approach to investigating roundabout capacity. Conventional methods aim to identify the capacities of each single entrance, without any consideration of mutual interactions between all the entrances and exits of the roundabout. This study investigated the interrelationships between traffic operations at the elements of a single-lane roundabout based on probabilistic considerations obtained from queueing theory. Here the spread of potential congestion on the circle plays a decisive role. As a starting point, the elements of the roundabout were defined as points of conflict. Equations for the capacity of these points were then derived. These equations were combined into an iterative procedure that took into account the potential congestion on the circle. As a result, based on a given origin–destination matrix of traffic volumes, the entry capacities of the roundabout were calculated. These were then applied to calculate estimates for delays and queue lengths in the conventional way. Application of the model for several examples—two of which are described in this paper—showed rather drastically that conventional evaluations of traffic performance at roundabouts may be misleading, with overestimated capacity and underestimated delays. This happens preferably in a range with volume/capacity values above 0.7 and under significant pedestrian traffic. The study underlines that the current assessment methods of analytical performance at roundabouts fall short.
Read full abstract