基于时空插值的海量车辆轨迹数据索引方法
首发时间:2021-02-05
摘要:随着定位和通信技术的发展,道路系统中很多车辆都安装了位置信息的采集、存储及传输设备,所获得的车辆轨迹大数据成为交通研究的重要基础。由于轨迹大数据具有海量性及时空异变性等特点,如何实现对其有效存储和应用成为当前研究的挑战和热点。本文就此提出了一种多时空粒度索引结构,称为"MSTGI",其可以在建立索引结构的同时对当前轨迹数据进行不同时间粒度的插值,并实现对轨迹信息的快速查询。该方法首先从原始轨迹数据中获取车辆标识符集合,然后将原始轨迹数据按车辆标识符组织成有序轨迹序列,并对该序列进行时间插值、空间插值及状态插值。本文以上海市约一万三千辆出租车连续一个月的轨迹数据为例,将该方法与已有索引方法进行了对比,结果表明,本文方法在保证索引构建效率的同时,有效改善了已有索引方法应用于轨迹大数据时存在的查询及存储效率问题。
For information in English, please click here
Mass Vehicle Trajectory Data Indexing Method Based On Spatiotemporal Interpolation
Abstract:With the development of positioning and communication technology, many vehicles in the road system have installed location information acquisition, storage and transmission equipment, and the obtained vehicle trajectory big data has become an important basis for traffic research. Due to the characteristics of trajectory big data with massive time and space variability, how to realize its effective storage and application has become a challenge and hotspot of current research. In this paper, a multi-time-space granular index structure called "MSTGI" is proposed, which can interpolate the current trajectory data at different time granularity while establishing the index structure, and realize fast query of the trajectory information. The method firstly obtains a set of vehicle identifiers from the original trajectory data. Then it organizes the original trajectory data into an ordered trajectory sequence by vehicle identifier, and performs time interpolation, spatial interpolation and state interpolation on the sequence. This paper takes the trajectory data of about 13,000 taxis in Shanghai for one month as an example. The method is compared with the existing index method. The results show that the proposed method can not only effectively improve the query and storage efficiency of existing indexing methods but also ensure the efficiency of index construction when applied to trajectory big data.
Keywords: Trajectory big data Spatiotemporal index structure Multiple spatiotemporal granularity index Spatiotemporal interpolation
引用
No.****
动态公开评议
共计0人参与
勘误表
基于时空插值的海量车辆轨迹数据索引方法
评论
全部评论0/1000