基于K-Means和FCM的增强型Wi-Fi指纹定位策略

摘要

The data processing algorithm is studied to improve Wi-Fi fingerprint indoor positioning performance. Firstly, Wi-Fi fingerprint samples are collected and then are put into MySQL database and R project. Secondly, the Wi-Fi fingerprint data is divided into several clusters, and the K-mean clustering (K-Means) and fuzzy C-means clustering (FCM) are used to cluster the Wi-Fi fingerprint respectively. Finally, an enhanced clustering strategy (ECS) is proposed to for Wi-Fi fingerprint matching. Experimental results show that ECS reduces the positioning time-consuming about 50%-80% than that consumed by only using FCM and the positioning accuracy is also improved; ECS improves about 20%-40% than that obtained by only using K-Means in terms of positioning accuracy and it proves positioning stability and can automatically update the Wi-Fi fingerprint database.%研究了通过数据处理算法以提高Wi-Fi指纹库室内定位性能的问题.首先采集Wi-Fi指纹样本,将其放入MySQL数据库中和R工程;其次将Wi-Fi指纹库分成若干个簇,使用K-均值聚类(K-Means)和模糊C-均值聚类(FCM)对待定位的Wi-Fi指纹进行聚类分析;最后,提出增强型的聚类策略(ECS)应用于Wi-Fi指纹匹配定位中.实验结果表明,ECS较仅使用FCM算法,其定位耗时缩短约50%-80%,且定位精度上有所改善;ECS较仅使用K-Means算法,其定位精度提高约20%-40%,且定位稳定性较强并自动更新Wi-Fi指纹库.