Analysis of Spatiotemporal Dependencies in Two-Dimensional Traffic Flow in Large-Scale Urban Area with Probe Vehicle Data

Abstract

Understanding and description of traffic dynamics in large-scale networks is difficult and challenging procedure due to complexity of the network, limited amount of available information regarding vehicle movements and great number of different factors which affect traffic conditions. To deal with these issues, in current study large-scale urban areas is substituted by a continuous medium, and traffic flows are treated as two-dimensional flow on it. Spatiotemporal characteristics of this two-dimensional flow were investigated with the help of actual large-scale probe vehicle data. Specifically, probe vehicle data, sampled from the vehicles travelled in Tokyo area for one month period were used. After data preprocessing and transformation, spatiotemporal dependencies in the whole area, as well as in smaller regions were examined. The results showed distinct patterns according to different spatial locations, temporal intervals and local features, such as existence of highways or high degree of arterials roads. These results suggested the applicability of proposed approach in order to analyze and describe dependencies in two-dimensional traffic flow by means of probe vehicle data.