摘要
目前已有的位置指纹室内定位算法大多都是建立在原始指纹库的基础之上,指纹库的建立精度会直接影响到最终的定位精度。为此,通过对指纹数据的研究,提出一种基于正态检验的室内定位算法。训练阶段,首先对每个指纹点接收到的信号RSSI样本进行正态假设检验,若接受假设则选用正态分布函数对其总体进行概率密度估计,否则选用核函数对其总体进行概率密度估计,最后取大概率信号的均值建立高精度的指纹数据库。在线定位阶段,使用K加权邻近算法(WKNN)估算位置,实验结果表明提出的算法定位精度较均值模型法以及正态模型法都提高了15%以上。
Currently,most existing fingerprinting indoor location algorithms are based on the original fingerprint database,the accuracy of establishing fingerprint database will directly affect final localization precision. Through the study of fingerprint data propose a database optimization algorithm based on the assumption of normality test. Training phase,the normal hypothesis test is performed on the RSSI samples of the signal which is received at each point of the fingerprint firstly. if accept the hypothesis,then use normal distribution function to do estimation of probability density,or use kernel function,finally select the mean value of sipnal with high probalility,to tuild the fingerprint database.During coline positioniy stage,use WRNN to estimate location. Experiment results show the proposed algorithm have higher accuracy.
出处
《激光杂志》
北大核心
2017年第3期41-45,共5页
Laser Journal
基金
江苏省第十一批"六大人才高峰"高层次人才项目
江苏省自然科学基金(BK20161536)
江苏高校优势学科Ⅱ期建设工程资助项目
关键词
室内定位
正态假设检验
正态分布
核函数
indoor positioning
normality assumption
normal distribution
kernel function
