当前位置: 首页> 国外交通期刊数据库 >详情
原文传递 A scenario-based map-matching algorithm for complex urban road network
题名: A scenario-based map-matching algorithm for complex urban road network
正文语种: 英文
作者: Xiangfu Kong; Jiawen Yang
作者单位: Shenzhen Graduate School, Peking University, Shenzhen, China
关键词: Complex road network; map-matching algorithm; taxi GPS data; trajectory
摘要: Previous map-matching algorithms perform well in sparse road areas and simple road network structures. Nevertheless, mismatches emerge when the algorithms are implemented in a complex road network. Special road patterns such as parallel or auxiliary roads, roundabouts, intersections, U-turns, or overpasses make it harder to identify features correctly. This paper proposes a scenario-based map-matching algorithm which balances accuracy and efficiency. Instead of matching every point by a constant process, the proposed algorithm provides specific combinations of filters for different operational environments. To eliminate improper candidate roads at each step, five road filters are introduced. Each filter has its own trigger condition, and if the condition is not satisfied, the filter will be skipped. The algorithm is validated using taxi GPS data on various complex roads in Shenzhen. The algorithm correctly identified 99.60%, 96.40%, 93.18%, 96.87%, and 94.11% of the roads for simple, opposing parallel, and auxiliary roads; intersections; and unconnected buffer situations, respectively. The proposed algorithm has two important features: 1. the algorithm is suitable for various complex intersections; 2.the map-matching process can be adjusted according to the traffic scenario, which saves computational time. The evaluated performance indicates that the proposed algorithm performs better than previous map-matching algorithms in both simple and complex road networks and can be efficiently applied to GPS datasets using various time intervals.
出版年: 2019
期刊名称: Journal of Intelligent Transportation Systems Technology Planning and Operations
卷: 23
期: 6
页码: 617-631
检索历史
应用推荐