In this paper, urban traffic is modeled using dual graph representation of urban transportation network where roads are mapped to nodes and intersections are mapped to links. The proposed model considers both the navigation of vehicles on the network
and the motion of vehicles along roads. The roads capacity and the vehicle-turning ability at intersections are naturally incorporated in the model. The overall capacity of the system can be quantified by a phase transition from free flow to congestion. Simulation results show that the systems capacity depends greatly on the topology of transportation networks. In general, a well-planned grid can hold more vehicles and its overall capacity is much larger than that of a growing scale-free network.
