作者: Oliver Jan , Alan J. Horowitz , Zhong-Ren Peng
DOI: 10.3141/1725-06
关键词: Street network 、 Telecommunications 、 Global Positioning System 、 Computer science 、 Shortest path problem 、 Statistics 、 Path (graph theory) 、 Set (abstract data type)
摘要: A comprehensive set of Global Positioning System (GPS) vehicle location data from Lexington, Kentucky, households was analyzed to determine if such can be helpful in improving path choice assumptions traffic assignment models. The portion the used consisted primarily a reconstruction street network and lists segments each path. Analysis based on matches trips (e.g., pairs with similar origins destinations). Matches were obtained for within across households. Statistics compare deviation index percentage identical links. It found that chosen trip quite sensitive origin destination most often differed considerably shortest time network. Paths made by same driver very consistent over time; paths different drivers showed more deviations even when ends or similar. As result this research, recommendations are as how GPS better collected future improvement