Abstract

This paper presents an indoor localization system based on Bisecting k-means (BKM). BKM is a more robust clustering algorithm compared to k-means. Specifically, BKM based indoor localization consists of two stages: offline stage and online positioning stage. In the offline stage, BKM is used to divide all the reference points into k clusters. A series of experiments have been made to show that our system can greatly improve localization accuracy.

中文翻译：

---

基于二等分k均值的指纹室内定位

摘要

本文提出了一种基于二等分k均值（BKM）的室内定位系统。与k均值相比，BKM是更健壮的聚类算法。具体地，基于BKM的室内定位包括两个阶段：离线阶段和在线定位阶段。在离线阶段，BKM用于将所有参考点划分为k个簇。进行了一系列实验，表明我们的系统可以大大提高定位精度。