基于GPS的浮动车数据与实地图匹配的算法研究

发布时间：2018-01-19 19:28

本文关键词： 浮动车地图匹配权重拓扑结构全球定位系统　出处：《武汉理工大学》2013年硕士论文　论文类型：学位论文

【摘要】：地图匹配(Map Matching)是一种广泛应用于导航系统定位的修正方法,这种方法是以软件技术为基础的,通过算法模型将GPS终端设备获取的车辆行驶轨迹与实地图中的拓扑道路关联起来,并修正车辆在实地图的相对行驶位置。由于地图匹配在交通系统中有相当大的实用性,很多学者花费大量的时间进行了研究,提出了很多与地图匹配相关的算法,这些算法各有优劣。能否减少导航系统定位的误差直接体现了地图匹配算法性能的好坏,但算法的实时性和鲁棒性也不容忽视。由于现有的地图匹配算法大多只考虑算法自身,导致算法实用性相对不强,所以应该考虑将其他可以有效利用的元素融合到算法中以提高算法的实用性。总而言之,在设计地图匹配算法的过程中,比较多种算法的优缺点,并将这些算法有效地结合起来,最终实现具有较高实时性、鲁棒性以及匹配精度的地图匹配算法。本文分析了影响地图匹配算法匹配精度的因素,并对现有的地图匹配算法进行了一系列的研究与总结,在此基础上,提出一个以道路网的拓扑特性为前提,利用当前车辆行驶位置的经纬度、行驶方向等信息与电子地图进行匹配的匹配算法。本论文提出的算法的基本思想是通过对电子地图进行网格划分来确定GPS定位点的网格编号,再通过构建误差区域来缩小匹配的搜索范围,从而减少了候选匹配路段数,最后构建了一个匹配模型,用来计算GPS定位点与候选匹配路段的匹配度,该模型即为权重模型。该权重模型是通过计算GPS定位点与候选匹配路段的最短距离、GPS定位点的行驶方向与候选匹配路段的夹角、定位点与候选匹配路段的相对位置关系这三个要素的权重总和来确定匹配路段,依据道路的拓扑关系、交通规则管制,在权重总和最大及次大的待匹配路段中选取最佳匹配路段。若仍无法从权重最大和次大的候选匹配路段中选取最优匹配路段,则将此定位点的相关信息记下,继续后续定位点的匹配,待后续定位点匹配完成后,利用最短路径算法来确定此定位点的匹配路段,再通过垂直投影法估计GPS定位点在此路段上的匹配位置。最后,在MATLAB平台上进行仿真实验,结果证明该算法是正确有效的。与传统的基于权重的地图匹配算法相比,改进后的基于权重的地图匹配算法通过网格划分和构建误差区域降低了在选取候选匹配路段时的时空复杂度。
[Abstract]:Map matching is a correction method widely used in navigation system location, which is based on software technology. Through the algorithm model, the vehicle track obtained by the GPS terminal device is associated with the topological road in the real map. Because map matching is very practical in traffic system, many scholars have spent a lot of time studying it. Many algorithms related to map matching are put forward, which have their own advantages and disadvantages. Whether the positioning error of navigation system can be reduced directly reflects the performance of map matching algorithm. However, the real-time and robustness of the algorithm can not be ignored. Because most of the existing map matching algorithms only consider the algorithm itself, the practicability of the algorithm is relatively weak. So we should consider merging other effective elements into the algorithm to improve the practicability of the algorithm. In short, in the design of map matching algorithm, compare the advantages and disadvantages of many algorithms. Finally, a map matching algorithm with high real-time, robustness and matching accuracy is realized by combining these algorithms effectively. This paper analyzes the factors that affect the matching accuracy of map matching algorithm, and carries on a series of research and summary to the existing map matching algorithm. On this basis, we propose a road network topology characteristics as the premise. Use the latitude and longitude of the current vehicle position. The basic idea of the algorithm proposed in this paper is to mesh the electronic map to determine the grid number of the GPS location point. Then by constructing the error region to reduce the search range of the matching, the number of candidate matching sections is reduced. Finally, a matching model is constructed to calculate the matching degree between GPS positioning points and candidate matching sections. The weight model is based on the calculation of the shortest distance between the GPS location point and the candidate matching section and the angle between the driving direction of the GPS location point and the candidate matching section. The relative position relationship between the location point and the candidate matching section is determined by the sum of the weights of the three elements. According to the topological relationship of the road, the traffic rules control. The best matching section is selected in the section to be matched with the largest and the second largest weights. If the optimal matching section can not be selected from the candidate matching section with the largest and the second largest weight, the relevant information of this location point is noted down. Continue the matching of the following location points, and then use the shortest path algorithm to determine the matching section of the location point after the completion of the subsequent location point matching. Then the matching position of GPS positioning point on this section is estimated by vertical projection method. Finally, the simulation experiment is carried out on the MATLAB platform. The results show that the algorithm is correct and effective, compared with the traditional weighted map matching algorithm. The improved weighted map matching algorithm reduces the space-time complexity of selecting candidate matching sections by mesh generation and error region construction.
【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：U495;P228.4

【参考文献】