摘要
综合采用全量详单(XDR)数据、互联网实时出行数据、隐马尔可夫模型(HMM)和维特比算法(Viterbi),提出一种新的指纹定位方法。在LTE号码回填和指纹定位基础上,首先在XDR数据特定OTT的URL中提取经纬度,然后结合互联网实时出行数据完成道路用户识别,最后利用隐马尔可夫模型和维特比算法完成用户定位修正和地图匹配,测试验证表明,新算法道路匹配率达到98%以上。
A new fingerprint localization algorithm is proposed by using XDR data,Real-time Internet travel data,HMM and Viterbi.Based on LTE number backfill and fingerprint location,firstly it extracts latitude and longitude data from URL of specific OTT of XDR data,and then combines the real-time travel data to complete the road user identification.Finally,using hidden Markov model and Viterbi algorithm,it completes user location correction and map matching.Test results show that the road matching rate of the new algorithm is over 98%.
作者
赵伟
王勇
刘宏嘉
Zhao Wei;Wang Yong;Liu Hongjia(China Unicom Zhejiang Branch,Hangzhou 325800,China;China United Network Communi cations Group Co.,Ltd.,Beijing 100033,China)
出处
《邮电设计技术》
2019年第11期20-23,共4页
Designing Techniques of Posts and Telecommunications
关键词
隐马尔科夫模型
维特比算法
指纹定位
MDT
Hidden markov model
Viterbi
Fingerprint location
Minimization of drive-tests
