A spatial analysis approach for describing spatial pattern of urban traffic state

To urban transportation management, it is valuable to obtain the whole road network traffic state, while most existing traffic analysis methods pay little attention to this problem. This paper proposes a method to analyze spatial pattern of road network traffic state, which describes traffic characteristic from network level. Above all, we treat traffic data as road link's attribution and establish a spatial model to describe road network traffic data from spatial perspective. Then using real travel speed data extracted from long-term floating car data, we analyze spatial dependency of traffic state based on spatial autocorrelation method, which shows that urban road network traffic state exist association in space, and this method can quantify the influence degree of road link, which is useful to micro-view traffic management; using trend surface analysis method, we analyze spatial heterogeneity of urban traffic state, which qualitatively shows that urban traffic state is unstable in space and is related to land use. Further, divergence tendency analysis shows that urban traffic state is influenced by road network structure, and it can also quantitatively analyze the influence range, which is useful to detailedly discover hot spot in urban traffic, such as traffic congestion. This study shows that it is a feasible approach to use spatial analysis method to study the overall characteristics of urban traffic system, which is of benefit to both dynamic traffic control and long-term traffic planning.