Keywords: Data independence, covariance matrix, traffic network
With over half the world's population now living in cities, resilience and efficiency in urban transportation networks are becoming ever more important. Modeling urban traffic networks may help answer the question of how to enhance transportation network’s resilience and efficiency, which has been an ongoing research topic in the field of transportation engineering. An important aspect of urban traffic network model is to ensure the independence of traffic links. In this paper, we propose a novel method for modeling urban traffic network in which we develop a technique of minimizing the dependence of link data (traffic condition). To be specific, we explore how to develop a general rule for choice of links in order to avoid dependence among the traffic flows. Note that we use Google Maps to collect traffic condition data for links. Then, after collecting independent link data, we apply the maximum likelihood estimation to estimate links travel times, and simulate the dynamic traffic network with a Markov framework.