Traffic flow behavior at a single lane traffic circle

In this paper, we propose a stochastic cellular automata (CA) model to study traffic dynamic at a single-lane traffic circle of N entry/exit points. The boundaries are controlled by the injecting rates α1, α2 and the extracting rate β. Both the cases with and without the splitter islands of width Lsp are considered. The phase diagrams in the space (α1β) are constructed and the density profiles are also investigated. The variation of the phase diagram with the traffic circle size (L), the number of entry point (N), and the time needed for drivers to change their exit willingly (Texit) is studied. The results show that the phase diagram in both cases consists essentially of three phases namely free flow, congestion and gridlock. However, the large sized traffic circle shows better performance in the free flow phase, in contrary, the increase of N enlarges the congestion and the gridlock phases. We have also found that the decrease of Texit enhances the traffic flow situation in the traffic circle. Furthermore, as the Lsp increases, the free flow phase enlarges while the congestion and gridlock ones shrink.