离散度WKNN位置指纹Wi-Fi定位算法

发布时间：2018-01-21 07:05

本文关键词： 无线定位位置指纹 Wi-Fi 接收信号强度指示离散度　出处：《哈尔滨工业大学学报》2017年05期 　论文类型：期刊论文

【摘要】：为改善加权K近邻位置指纹定位算法在室内环境复杂时的定位性能,提出一种以位置指纹离散度作为权值参考的改进加权K近邻位置指纹定位算法.算法在离线位置指纹数据库建立阶段采用K-means聚类算法对位置指纹进行聚类,来降低搜索位置指纹库的计算量.从离线位置指纹库中选取K个与在线实测Wi-Fi信号强度信息最相似的位置指纹,比较其离散程度,将离散程度小的位置指纹赋予较高的加权系数,以减小原加权K近邻算法在室内复杂环境信号强度随距离变化较大情况下带来的位置估算误差.对离散度加权K近邻算法时间复杂度的分析表明,其计算量小于原加权K近邻算法;实际环境实验结果表明,离散度加权K近邻算法具有更高的定位精度,且定位误差波动较小.
[Abstract]:In order to improve the location performance of weighted K-nearest neighbor location algorithm in complex indoor environment. An improved weighted K-nearest neighbor location algorithm based on location fingerprint dispersion is proposed. K-means clustering algorithm is used in the establishment of off-line location fingerprint database. Row clustering. From the off-line location fingerprint database, we select K position fingerprint which is the most similar to the on-line measured Wi-Fi signal intensity information, and compare its discretization degree. The small degree of dispersion of the location fingerprint is given a higher weighting coefficient. In order to reduce the position estimation error caused by the original weighted K-nearest neighbor algorithm when the signal intensity of indoor complex environment varies greatly with the distance, the time complexity of the discrete-weighted K-nearest neighbor algorithm is analyzed. The computational complexity of the algorithm is less than that of the original weighted K-nearest neighbor algorithm. The experimental results show that the discretization weighted K-nearest neighbor algorithm has higher localization accuracy and less fluctuation of location error.
【作者单位】：吉林大学通信工程学院;东北电力大学信息工程学院;长春理工大学电子信息工程学院;
【基金】：国家自然科学基金(61371092,61401175,61540022) 吉林大学研究生创新基金(2016091)
【分类号】：TN92
【正文快照】： 随着网络技术的发展,无数智能节点通过无线通信技术连接在一起,形成了无所不能的物联网[1].其中位置服务[1-2]是物联网面向用户的关键需求之一,而位置服务主要依赖于无线定位技术.无线定位技术从地理位置上可分为室内定位以及室外定位.室外定位一般通过卫星导航系统辅助以AGPS

【相似文献】