基于中心矩改进的WiFi室内定位算法

IMPROVED WIFI INDOOR LOCALIZATION ALGORITHM BASED ON CENTRAL MOMENT

基于WiFi指纹的定位技术是国内外室内定位领域研究的热门课题。针对室内环境的实变性而造成RSS值波动带来的影响,提出一种基于中心矩加权改进的WKNN匹配算法。离线阶段中,提取RSS的均值和二阶矩作为指纹存入数据库。在线阶段定位时,将RSS值的二阶矩特征加入到欧式距离中进行计算,以得到离定位点最近的K个参考点,从而计算出待定位点的位置。实验结果表明:不同的匹配算法的选择会使定位误差呈现出明显的差异性,所提出改进的WKNN算法能够有效地提高室内定位精度。

Positioning technology based on WiFi fingerprint is a hot field in the research of indoor positioning at home and abroad. In terms of the impact of the RSS value fluctuation caused by the real degeneration in indoor environment,the paper proposes an improved WKNN matching algorithm based on weighted central moment. In the offline phase,we extract the mean and the second-order central moment of the RSS as a fingerprint database,in online phase,we add the second moment characteristics of RSS values to the Euclidean distance to calculate in order to get the nearest K reference points which are close to the location point,and calculating the position to be positioned. The experimental results show that the selection of different matching algorithms will make the positioning error show obvious differences,and the improved WKNN algorithm we propose can effectively improve the indoor positioning accuracy.