Modeling traffic flow at a single-lane urban roundabout

Abstract In this paper, we propose a new model to study traffic flow at a single-lane urban roundabout, using a multi-state cellular automata (CA) ring under the offside-priority rule (by which a vehicle entering gives way to one already on the roundabout). Each vehicle entering the roundabout is randomly characterized by a predetermined exit with specified probability. Driver behavior at the roundabout entrance is randomly grouped into four categories based on space required to enter the roundabout. Three aspects of roundabout performance in particular have been studied. The first looks at overall throughput (the number of vehicles that navigate the roundabout in a given time). This is considered for different geometries, turning and arrival rates (vehicles arrive at random with a Poisson distribution, with parameter λ ⩽0.5 in general for free flow). The second investigates changes in queue length, delay time and vehicle density (ratio of the number vehicles to the number of cells) for an individual road. The third considers the impact of driver choices on throughput and operation of the roundabout. We find that throughput is influenced by the topology of the roundabout and turning rates, but only incidentally by size. Throughput reaches a maximum for critical arrival rate on one or more roads. Driver behavior has considerable impact on overall performance, with rapid congestion resulting from reckless choices. Vehicles drive on the left in Ireland, but rules are generally applicable.