摘要
针对基于加权K最近邻(WKNN)的WiFi指纹定位算法精度低的问题,提出了基于归一化接收信号强度(RSS)和约束WKNN的WiFi指纹定位算法。采用高斯滤波对离线阶段和在线阶段采集的RSS值去噪,降低信号的随机误差,并建立位置指纹库(radio map);采用基于4—域系统的WKNN算法匹配定位,防止离待测点较远的参考点参与匹配造成的误差。实验结果表明:改进后的WiFi指纹定位算法可以更好地估计用户的实际位置,平均定位误差降低了19.4%。
Aiming at the problem of low positioning precision of WiFi fingerprint positioning algorithm based on weighted K-nearest neighbor（ WKNN）,a WiFi fingerprint localization algorithm based on the normalized received signal strength（ RSS） and constraint WKNN is proposed. Use Gaussian filtering for denoising on RSS value acquired in offline and online stage,decrease the random error of signal,and construct radio map; WKNN algorithm based on 4-domain system is used to match and locate,to prevent the reference points far from the specific points to participate in the match,so as to results in error. The experimental results show that the improved WiFi fingerprint positioning algorithm can better estimate the user ＇s actual location,the average positioning error is reduced by 19. 4 %.
作者
冯涛
阮超
郭凯旋
卢彦霖
余敏
FENG Tao;RUAN Chao;GUO Kai-xuan;LU Yan-lin;YU Min(School of Computer Information and Engineering,Jiangxi Normal University,Nanehang 330022,China;School of Software,Jiangxi Normal University,Nanchang 330022,China)
出处
《传感器与微系统》
CSCD
2018年第10期127-129,共3页
Transducer and Microsystem Technologies
基金
国家自然科学基金资助项目(41374039)
国家重点研发计划课题资助项目(2016YFB0502204)
关键词
归一化
高斯滤波
加权K最近邻
4—域系统
WiFi指纹
normalization
Gaussian filtering
weighted K-nearest neighbor （WKNN）
4-domain system
WiFi fingerprint
