基于Hadoop云平台的矿井指纹定位算法  

Mine Fingerprint Localization Algorithm Based on Hadoop Cloud Platform

作　　者：韩继英[1] 李芳丽 朱元凯[3] HAN Jiying;LI Fangli;ZHU Yuankai(Department of Computer Engineering,Shanxi Vocational and Technical College,Taiyuan 030006,China;School of Engineering,Lincoln University,Petaling Jaya 47301,Malaysia;Department of Information Technology Engineering,Taishan Vocational and Technical College,Tai′an 271001,China)

机构地区：[1]山西职业技术学院计算机工程系,山西太原030006 [2]马来西亚林肯大学工程学院,雪兰莪州八打灵再也47301 [3]泰山职业技术学院信息技术工程系,山东泰安271001

出　　处：《金属矿山》2023年第9期193-198,共6页Metal Mine

基　　金：江西省教育厅科技项目(GJJ180975)。

摘　　要：针对现有矿井指纹定位算法定位精度不高、实时性不强的问题,结合Hadoop云平台技术,提出了一种基于Hadoop云平台的矿井指纹定位算法。该算法首先在指纹特征提取前对矿井中的多基站信号源进行数据预处理,提高指纹特征的可靠性和鲁棒性;其次,利用小波基函数对源指纹信号进行线性变换,并将多个基站信号分解成不同尺度的频带,得到具有代表性的指纹特征向量;再次,将每个待定位的矿工或移动设备信号分解为若干个子信号区域,并逐区域计算待测信号源与特征集之间的相关系数矩阵,根据相关系数矩阵实现指纹信息匹配;最后,利用多普勒效应和信号衰减原理,对每个节点的位置进行估计,获得每个节点的具体位置信息。在国内某矿井进行了定位试验,结果表明:相对于主流的指纹定位算法,所提出的矿井指纹定位算法能够快速、准确地定位矿井中的人或移动设备,具有较高的定位精度和实时性。该算法可以应用于实际的矿井安全监测和管理中,为矿井安全和高效生产提供支持。Aiming at the current problems of the low positioning accuracy and weak real-time performance of existing mine fingerprint localization algorithms,a Hadoop cloud platform-based mine fingerprint localization algorithm is proposed.The proposed algorithm first preprocesses the multi-base station signal sources in the mine before fingerprint feature extraction to improve the reliability and robustness of the fingerprint features.Secondly,the source fingerprint signal is transformed linearly using wavelet basis functions,and the signals from multiple base stations are decomposed into frequency bands of different scales to obtain representative fingerprint feature vectors.Furthermore,for each location-based signal of the miner or mobile de-vice,it is decomposed into several sub-signal regions,and the correlation coefficient matrix between the test signal source and the feature set is calculated for each region,based on which matching of fingerprint information is implemented.Finally,using the doppler effect and signal attenuation principle,estimate the position of each node to obtain specific location information for each node.A localization experiment is conducted on a mine in China,and the results showed that compared with mainstream fingerprint localization algorithms,the proposed mine fingerprint localization algorithm can quickly and accurately locate people or mobile devices in the mine,with high positioning accuracy and real-time performance.The proposed algorithm can be applied to practical mine safety monitoring and management,providing support for mine safety and efficient production.

关 键 词：矿井指纹定位 HADOOP 云平台 MAPREDUCE 指纹匹配 

分 类 号：TD67[矿业工程—矿山机电]