Spatial Autocorrelation of Urban Road Traffic Based on Road Network Characterization

Abstract

Urban road traffic is spatially autocorrelated.The change of traffic on certain road will quickly affect the traffic on nearby roads,which will alter the overall traffic status within a neighborhood.Revealing the spatial autocorrelation structure in urban road traffic is important for traffic planning,traffic controlling and traffic guidance.The traffic interaction between neighboring roads is not isotropy.The traffic change on certain road does not equally spread to each spatially adjacent road,but concentrate on some of them.Thus only using spatial adjacency to define adjacent roads cannot well reveal the spatial autocorrelation in urban road traffic.Recent research has proved that the dynamic flow on networks highly depend on the structure of networks.Characterizing the structure of urban road network is essential to reveal the spatial autocorrelation in urban road traffic.The aim of this research is to reveal the spatial autocorrelation of urban road traffic based on road network characterization.We first investigate the modular character and hierarchal feature of urban road network quantitatively.The modulars in road network are defined as a group of closely connected neighboring road segments and identified by community detection algorithm from complex network theory.The hierarchal feature of urban road network helps to determine the structural importance of road segments.Topological roles are defined based on the structural importance of road segment.Then we provide a novel approach to define adjacent road segments based on the topological roles in spatially adjacent road segments.Two road segments defined as adjacent road segments not only locate in a nearby neighborhood but also have the same topological roles.A set of adjacent roads constitute a spatial related set.Experiment results on the road network of Beijing imply that the spatial related sets identified by the proposed approach can capture the spatial autocorrelation structure of urban road traffic.