电子测量与仪器学报

米伟娟,李娜.基于变分自编码器的 WLAN 定位方法[J].电子测量与仪器学报,2020,34(12):101-108

基于变分自编码器的 WLAN 定位方法

Variational auto encoder based WLAN positioning

DOI：

英文关键词:variational auto encoder WLAN based positioning neural network

基金项目:

作者	单位
米伟娟	1.河北传媒学院信息技术与文化管理学院
李娜	1.河北传媒学院信息技术与文化管理学院

Author	Institution
Mi Weijuan	1.Information Technology and Cultural Management Institue, Hebei Institute of Communication
Li Na	1.Information Technology and Cultural Management Institue, Hebei Institute of Communication

摘要点击次数: 311

全文下载次数: 304

中文摘要:

基于无线局域网络(WLAN)的定位是当前室内定位领域的主流技术之一。提出一种基于变分自编码器的 WLAN 定位方法,具备定位通路和接受信号强度(RSS)信号指纹重构通路,使得该方法既具备定位能力,也具备重新生成 RSS 信号指纹或指纹地图的能力。利用开源的数据集进行验证,证明了该方法在定位通路上,相比稀疏重构定位方法和传统 k 最邻近( kNN)定位方法,误差分别下降了约 14%和 24%;能够对指纹地图(RM)进行重新生成,利用重构的 RM 进行定位,相比于利用稀疏重构生成的 RM 定位误差下降了约 11%。

英文摘要:

Wireless local area network (WLAN) based positioningis one of the main stream techniques for indoor positioning. In this paper, a WLAN based method is proposed based on a variational auto encoder ( VAE), which contains a positioning path and a reconstruction path. This structure has enabled the ability to position along with reconstruction of RSS radio map (RM). The proposed methodis validated by open-sourced dataset, which shows that the proposed method can increase the positioning accuracy by about 14% and 24% respectively compared to the compressive sensing (CS) based methodand the traditional kNN method in the positioning path. Moreover,from the reconstructed RM, the method can have accuracy enhancement of about 11% in positioning than that from the CS reconstructed RM. Keywords:variational

查看全文查看/发表评论下载PDF阅读器