On Identifying Dynamic Intersections in Large Cities

As the complexity of traffic conditions in large cities increases it becomes important and highly desirable to be able to analyse the spatial and temporal nature of such systems. Due to the heterogeneity of traffic demand and road network topology, it is possible to find 'hot spots' in a network that present a challenge for city planning and conventional intersection control methods. This paper presents an approach to identify such places as physical intersections in a traffic network with dynamically changing demand conditions in time and to quantify the level of volatility at those locations. We design a model that is used to simulate commuters path choices using a stochastic routing approach. We perform a case study for the city of Singapore and calibrate our model with national survey data describing the travel habits of the population. The results from our simulation are used to analyse the traffic conditions in the city. We are able to identify and study highly dynamic intersections and observe that such locations in fact exist and contribute to the heterogeneous dynamic profile of the road network.