Weight-Based Shortest-Path Aided Map-Matching Algorithm for Low-Frequency Positioning Data
Weight-Based Shortest-Path Aided Map-Matching Algorithm for Low-Frequency Positioning Data
复制标题
DOI：
--
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
Yuheng Zheng;M. Quddus
中科院分区：
文献类型：
--
作者：
Yuheng Zheng;M. Quddus
利用道路段连接性沿着与其它数据（即，位置、速度和航向）的现有地图匹配算法主要适用于高频（1Hz或更高）定位数据。当将这些算法应用于低频数据（诸如来自公共汽车或轻型车辆的车队的数据）时，这些算法的性能在正确的链路识别方面显著降低。这样的性能可能不适合于一些实时智能交通系统（ITS）的应用，如估计从低频GPS数据的链路行程时间。因此，本文提出了一种基于权重的最短路径辅助地图匹配（spMM）算法，以增强低频数据的地图匹配过程。著名的A* 搜索算法，同时考虑到链路连接和路口转弯限制，推导出两个连续的固定之间的最短路径。在开发的spMM算法中，两个额外的权重相关的最短路径沿着与两个权重（即接近度和方位差）通常使用在现有的地图匹配算法。一个附加权重与沿最短路径的距离沿着和沿车辆轨迹的距离沿着有关。另一个附加权重与从最短路径和车辆轨迹导出的航向信息相关联。所开发的spMM算法已经使用一系列不同频率（即1Hz，0.2Hz，0.033Hz和0.0167Hz）的真实世界数据集进行了测试。高精度组合导航系统（一个高级别的惯性导航系统和载波相位GPS接收机）被用来测试所开发的算法的性能。结果表明，spMM算法识别97.5%的链接正确的所有频率。在没有最短路径信息的情况下，该算法的性能降低到正确识别链路的70%。实验结果也表明该算法适合于实时应用。
Existing map-matching algorithms that utilize road segment connectivity along with other data (i.e. position, speed and heading) are primarily suitable for high frequency (1Hz or higher) positioning data. While applying these algorithms to low frequency data such as data from a fleet of buses or light duty vehicles, the performance of these algorithms reduces significantly in terms of correct link identification. Such a performance may not be suitable for some real-time Intelligent Transport System (ITS) applications such as estimating link travel time from low frequency GPS data. Therefore, this paper develops a weight-based shortest path aided map-matching (spMM) algorithm that enhances the map-matching process of low frequency data. The well-known A* search algorithm is employed to derive the shortest path between two consecutive fixes while considering link connectivity and turn restriction at junctions. In the developed spMM algorithm, two additional weights related to the shortest path along with the two weights (i.e. proximity and bearing difference) commonly used in existing map-matching algorithms are employed. One additional weight is related to the distance along the shortest path and the distance along the vehicle trajectory. The other additional weight is associated with the heading information derived from the shortest path and the vehicle trajectory. The developed spMM algorithm has been tested using a series of real-world dataset of varying frequencies (i.e. 1Hz, 0.2Hz, 0.033Hz and 0.0167Hz). A high accuracy integrated navigation system (a high-grade INS and a carrier-phase GPS receiver) has used to measure the performance of the developed algorithm. The results suggest that the spMM algorithm identifies 97.5% of the links correctly for all frequencies. Without the information from the shortest path, the performance of the algorithm reduces to in the region of 70% in terms of correct link identification. The results also suggest that the algorithm is suitable for real-time applications.