车辆导航系统中复杂路网的地图匹配技术研究

发布时间：2018-06-03 20:11

本文选题：车辆导航系统 + 地图匹配　；参考：《解放军信息工程大学》2014年硕士论文

【摘要】：当前车辆导航系统获得了广泛应用,普遍采用地图匹配技术解决卫星定位和制图误差等导致行车轨迹不在道路上的问题。地图匹配是借助电子地图对接收到的定位数据进行实时修正的过程。它能够减小定位误差对导航定位的影响,提高系统的定位精度。现有的地图匹配算法精度高、效果好,在大多数车辆导航系统中获得了较好的应用。然而,如果卫星信号被遮挡时导航系统不能在隧道、桥下有效定位,汽车就不能在复杂的立交桥上有效区分桥上桥下的情况。同时,对于应用复杂的军用车辆导航系统而言,除了导航应用外,它还包括复杂的地图数据处理、各种业务数据处理等功能。并且,较复杂的匹配计算在匹配区域切换时可能存在停顿,因而对匹配算法提出了更高要求。本文针对实际应用需求,对地图匹配算法的实现过程进行了深入研究,设计了一种具有强实时性,并对复杂路网匹配效果良好的地图匹配算法,论文的主要工作如下:(1)针对匹配计算时路网数据重组时间过长导致匹配停顿或延迟的问题,在基于优先级调度的多任务环境下采用“空间换时间”策略,实现了双缓存机制的路网数据访问方法。该方法首先在系统启动时申请了两块缓存空间,并采用基于节点-路段关联的匹配路网模型在前台缓存中组织路网数据。其次,建立高低任务优先级的主辅任务,高优先级的主任务在前台缓存中基于节点-路段关联模型进行实时匹配计算,而低优先级的辅任务则对后台缓存读取的区域路网数据进行预处理,有效减少了匹配搜索次数。同时,预先建立了新位置的节点-路段关联关系,避免了新位置的匹配切换引起的停顿问题。性能分析表明,双缓存策略实现了地图匹配的流畅运行,提高了路网数据的访问效率,保证了车辆的实时匹配。(2)针对嵌入式环境中匹配路网搜索时,路段形状点数量与搜索次数过多影响匹配实时性的问题,研究了基于匹配误差的自适应限差矢量数据压缩预处理方法。该方法在地图匹配过程中增加了带有反馈环节的矢量数据压缩模块,并应用改进的自适应限差Douglas-Peucker算法对区域路网数据进行压缩预处理。改进算法分别以地图精度、道路间距及匹配误差的最大似然估计值为输入参数,赋予适当的权值后将输出的加权平均值作为精度限差对路网数据进行矢量压缩。实验结果表明,与压缩前相比,压缩后的路段点数明显减少,说明路网预处理能够有效降低路网数据量,加快搜索速度,从而提高地图匹配的实时性。(3)针对二维地图匹配算法对立交桥、隧道等立体化路段的匹配精度相对较低的问题,设计了基于高程信息的改进D-S证据理论地图匹配算法。该算法首先根据有无历史定位信息采用不同的搜索方式确定候选道路集;其次,为了引入高程信息的作用,通过给出的自定义高程函数在位置信息证据及方向信息证据融合后的联合支持度函数中增加了一个高程因子,之后依据改进后的联合支持度函数结果判定匹配路段;最后,针对有效定位数据及缺失定位数据,分别采用正交投影法及推算定位法对定位结果进行修正。实验结果表明,相比传统的D-S证据理论法,改进算法有效提高了立交桥区域的匹配准确率,改善了匹配算法的处理性能。(4)在课题组开发的××导航定位系统中设计并实现了基于注册机制的构件化地图匹配组件,经过测试与验证,该构件能够满足系统的应用需求。
[Abstract]:The current vehicle navigation system has been widely used. The map matching technology is widely used to solve the problem that the path of the vehicle is not on the road. The map matching is the process of real-time correction of the received location data with the aid of electronic map. It can reduce the influence of the positioning error on the navigation and positioning. The existing map matching algorithm has high precision and good effect, and it has obtained good application in most vehicle navigation systems. However, if the navigation system can not be effectively located under the tunnel and under the bridge when the satellite signal is blocked, the vehicle can not effectively distinguish the bridge under the complex bridge. At the same time, In the application of complex military vehicle navigation system, besides the navigation application, it also includes complex map data processing, various business data processing and other functions. Moreover, more complex matching computation may have a pause when switching in the matching area. Thus, a higher requirement for the matching algorithm is proposed. The realization process of graph matching algorithm is deeply studied, and a map matching algorithm with strong real-time performance and good matching effect for complex road network is designed. The main work of this paper is as follows: (1) the problem of matching stop or delay with the long reconfiguration time of the road network data in the matching calculation, and the multi task based on priority scheduling. In the environment, the method of "space time for time" is adopted to realize the access method of road network data in double cache mechanism. This method first applies two blocks of cache space when the system starts, and uses a matching road network model based on node link link to organize road network data in the front cache. The high priority main task is based on the node link association model in the front desk for real-time matching calculation, while the low priority auxiliary task preprocesses the regional road network data that the background cache reads, effectively reducing the number of matching search. At the same time, the new bit link link relationship is established in advance to avoid the new location. The performance analysis shows that the dual cache strategy realizes the smooth operation of map matching, improves the access efficiency of road network data, and ensures the real-time matching of vehicles. (2) the problem of real-time matching of the number of road shape points and too much search times when searching the road network in the embedded environment, The adaptive limit vector data compression preprocessing method based on the matching error is studied. In the process of map matching, the vector data compression module with feedback link is added, and the improved adaptive limit difference Douglas-Peucker algorithm is used to compress the regional road network data. The improved algorithm is based on the map accuracy, The maximum likelihood estimation of road distance and matching error is the input parameter, and the weighted average value of the output is used as the precision limit to compress the vector of the road network data. The experimental results show that the number of road points after compression is obviously reduced compared with the compression, which indicates that the road network preprocessing can effectively reduce the amount of road network data. To speed up the search speed and improve the real time of map matching. (3) aiming at the problem that the matching precision of the two dimensional map matching algorithm is relatively low in the stereo sections such as overpass and tunnel, the improved D-S evidence theory map matching algorithm based on elevation information is designed. Secondly, in order to introduce the candidate road set, in order to introduce the function of elevation information, a elevation factor is added to the joint support function of the position information evidence and the direction information evidence fusion, and then the matching section is determined according to the improved joint support degree function. Finally, the target is determined. Effective positioning data and missing location data are corrected by orthogonal projection method and calculation location method respectively. The experimental results show that compared with the traditional D-S evidence theory, the improved algorithm improves the matching accuracy of the overpass area effectively and improves the processing performance of the matching algorithm. (4) the x x navigation developed in the project group In the positioning system, a component-based map matching component based on registration mechanism is designed and implemented. After testing and verification, the component can meet the application requirements of the system.
【学位授予单位】：解放军信息工程大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495;U463.67

【参考文献】