摘要
由于无线信号在传播过程中受到反射、衍射、多经效应和天线方向等的影响,使用传统的定位算法并不能达到很好的精度。实验数据显示RSS值(接收信号强度)与物理距离之间存在多对多的现象,针对这一固有特性,提出了n元距离组的定位算法。在数据集的获取过程中采用高斯滤波对RSS值进行优化处理,通过计算待定位点与数据集中的RSS值的相似性得到映射距离集合;为了减小干扰点的影响,对距离集合使用K-means聚类算法进行分类。实验结果表明,该算法能有效地减小环境的影响,达到较好的定位精度,比传统算法更优越。
As the wireless signal is affected by reflection, diffraction, multipath effect and the antenna direction in propagation process, traditional position algorithm cannot achieve well accuracy. The experimental data show the many-many relationship between the RSS value and the physical distance. Responsing to this inherent characteristic, an algorithm based on n-tuple distance sets is proposed. Gaussian filter is used to optimize the value of RSS in data sets training by calculating the similarity of RSS between blind and datasets to obtain the mapping distance sets. In order to reduce the impact of interference points, K-means clustering algorithm is used to classify. Experimental results show that the algorithm can effectively reduce the environmental impact and achieve better positioning accuracy superior than the traditional algorithm.
出处
《计算机工程与设计》
CSCD
北大核心
2014年第5期1557-1561,共5页
Computer Engineering and Design
关键词
定位
接收信号强度
n元距离组
映射
相似性
干扰点
positioning
RSS
n-tuple distance sets
mapping
similarity
interference points
