基于改进模糊核聚类的室内定位方法研究

摘要

An indoor positioning method based on kernel fuzzy C-means (KFCM) and weighted K-nearest neighbor (WKNN) is proposed to reduce the positioning time and improve the positioning accuracy.Firstly,the number of clusters and the initial cluster center are determined by clustering by fast search and find of density peaks (CFSFDP) to overcome instability,and then WKNN is used for location matching to improve accuracy.The experimental results show that compared with non-clustering indoor localization,the proposed method can reduce the computation time with satisfied accuracy.In addition,compared with the method based on K-means,KFCM and CFSFDP,the proposed method has better clustering effect and positioning accuracy.%针对室内定位中,WiFi位置指纹法存在的定位实时性和精度的问题,提出一种基于改进模糊核聚类(KFCM)和加权K近邻(WKNN)结合的室内定位方法,旨在降低定位时间和改善定位精度.首先利用快速搜索和发现峰值聚类(CFSFDP)确定聚类数目和初始聚类中心,克服KFCM算法对初始聚类中心选取的依赖性而导致聚类结果不稳定的缺点,在此基础上,采用WKNN进行定位匹配,提高定位精度.实验表明,所提出方法相较于无聚类的室内定位方法,能在保证一定精度的前提下,减少定位计算量和时间.此外,将所提出方法与基于K均值、KFCM和CFSFDP的方法进行实验对比,结果显示,该方法具有更好的聚类效果和定位精度.