Optimization of Dynamic Speed-Homogeneous Space Using Location-Based Vehicle Data

Abstract

Road extent compared with the number of automobiles registered in South Korea is the lowest among OECD (Organization for Economic Cooperation and Development) countries. Currently, traffic information is provided based on nodes and links using the centerlines of the roads. However, detailed speed information is unavailable. This study presents an optimization method for space to represent a dynamic speed-homogeneous space. Vehicle directions were separated by route using azimuth data based on global navigation satellite system (GNSS) location information, and the time mean speed and space mean speed were used to determine the homogeneous space of vehicle speed through the gradual separation of space. The processing speed of large-capacity data was increased by using Geohash, which applied a quaternary basis segmentation method that split data into two directions of longitude and latitude. The numbers of homogeneous speed spaces along the two directions of the Gyeongbu Expressway were determined to be 127 and 170, respectively, by the 12-step segmentation method. Spaces not included in the range of homogeneous speed spaces amounted to 5.19% and 7.51%, respectively, along the northbound and southbound roadways of the Gyeongbu Expressway. These non-convergence spaces were analyzed by the influence of vehicles entering or exiting slowly from the merging and diverging areas of the freeway junctions and interchanges. In other words, these were analyzed by the difference of speed by lane and not the driving direction.