基于K-means聚类的室内三维定位算法  

Indoor 3D positioning algorithm based on K-means

作　　者：周满满 袁凌云[2] ZHOU Man-man;YUAN Ling-yun(Key Laboratory of Educational Information for Nationalities of Ministry of Education,Yunnan Normal University,Kunming 650500,China;School of Information Science and Technology,Yunnan Normal University,Kunming 650500,China)

机构地区：[1]云南师范大学民族教育信息化教育部重点实验室,云南昆明650500 [2]云南师范大学信息学院,云南昆明650500

出　　处：《计算机工程与设计》2019年第9期2530-2536,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(61561055);教育部人文社会科学研究青年基金项目(13YJCZH223);云南省应用基础研究计划面上基金项目(2013)

摘　　要：为解决室内定位误差大、难实现三维定位的难题,提出一种基于K-means聚类的三维定位算法,对指纹数据库处理并同时实现三维定位,将RSSI相关系数与传统算法融合加权,使每个位置上的RSSI值直接映射出该位置的位置信息。实验结果表明,该算法能成功实现三维定位,提高平面定位精度,相对传统算法,将定位准确率提高了9.23个百分点,达到91.13%,将误差减少了95.81 cm,将定位误差控制在93.03 cm以内。The indoor positioning errors are large,and it is difficult to achieve 3D positioning,to solve the problems,a kind of three-dimensional positioning based on K-means clustering algorithm was put forward.The fingerprint database was processed while 3D positioning was achieved at the same time,then the RSSI correlation coefficient was weighted fused with the traditional method,the RSSI value of each position directly reflected the position information of the location.Experimental results show that the proposed algorithm can successfully realize 3D positioning and improve the accuracy of planar positioning.Compared with the traditional algorithm,the positioning accuracy is increased by 9.23 percentage points to 91.13%,the error is reduced by 95.81 cm,and the positioning error is controlled within 93.03 cm.

关 键 词：三维定位 室内定位 K-MEANS聚类算法 相似度权重量化 位置指纹 接收信号强度 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]