摘要
提出1种基于网格划分的轨迹相似性匹配技术,能快速高效地应用于海量车辆GPS数据与路网的匹配。首先以实际交叉口为节点构建泰森多边形,将路网划分为子区域;其次,在子区域,按照拓扑关系计算出路网节点之间的所有有效路径轨迹,以欧式距离作为相似度指标,对有效路径轨迹和车辆GPS轨迹进行相似度测算。该方法大大简化最短路径的重复搜索和计算,同时按照拓扑规则增加了轨迹匹配的方向性约束。本文已应用在某市连续3个月的出租车GPS匹配当中,为城市交通改善提供支撑。
Vehicle GPS trajectory data is an important data base for optimizing urban traffic operation,establishing traffic models,and updating and improving traffic facilities.This paper proposes a trajectory similarity matching technique based on grid division,which can be quickly and efficiently applied to match massive vehicle GPS data with road network.Firstly,the Tyson polygon is constructed with the actual intersection as the node,and the road network is divided into sub-regions;secondly,in the sub-regions,all the effective path trajectories between the nodes of the road network are computed according to the topological relationship,and the Euclidean Distance is used as the similarity index to measure the similarity between the effective path trajectories and vehicle GPS trajectories.This method greatly simplifies the repeated search and calculation of the shortest path,and at the same time increases the directional constraints of trajectory matching according to topological rules.This paper is applied to the GPS matching of cabs in a city for three consecutive months to provide support for transportation improvement.
作者
陈仲
CHEN Zhong(China Academy of Urban Planning and Design,Beijing 100037,China)
出处
《交通工程》
2024年第7期54-59,共6页
Journal of Transportation Engineering
