一种新的智能地图匹配算法  

New intelligent map-matching algorithm

作　　者：马建军[1] 唐进君[2] 曹凯[2] 

机构地区：[1]天津大学管理学院,天津300072 [2]山东理工大学交通与车辆工程学院,山东淄博255049

出　　处：《计算机应用》2007年第12期3116-3118,共3页journal of Computer Applications

摘　　要：为克服当前定位点信息不足的缺点,充分利用导航定位的历史轨迹信息,在分析常见地图匹配算法的基础上,引入了Fréchet距离来定义两曲线间的距离,并且通过设计一种智能云模型控制器对地图匹配这种具有高度不确定性的算法进行了云规则推理,最后推导出可信度P作为地图匹配效果的评价指标。该算法不仅能够在出现匹配错误时为使用者提供警告信息,而且还提供了一种能迅速从错误中调整恢复的方法。实验表明,当充分利用当前定位点信息并与历史轨迹信息结合时,匹配的总体精度要优于只利用当前定位点信息的情况,并且算法的准确性也比后者高。To overcome the disadvantage of inadequate information in map-matching algorithm for using current positioning point only, an intelligent map-matching algorithm was proposed by utilizing both current positioning point and some history positioning points. In this algorithm, an average Fréchet distance was employed to define the distance between two history positioning trajectories. Additionally, a credibility, P, as an evaluation index was generated by cloud controller so that the proposed algorithm could provide not only warning information to user when mismatching occurred, but also a means of fast recovery from failure. The simulation test demonstrates that the precision and accuracy of the proposed algorithm is superior to that of matching algorithm only using the current positioning point.

关 键 词：导航定位 地图匹配算法 Fréchet距离 云模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP301.6[自动化与计算机技术—计算机科学与技术]