摘要 WLAN（Wireless Local Area Networks）室内定位已受到人们广泛的关注，而在离线指纹采集阶段常常容易造成位置指纹RSS数据采集的盲目性和不可靠性，并忽略所需采集RSS（Received Signal Strength）样本容量与定位性能的关系.为了解决这一问题，本文提出一种面向WLAN室内定位的T检验样本容量优化方法.该方法在离线阶段利用OC（Operating Characteristics）函数优化指纹数据库允许的最小RSS样本容量，而在在线阶段利用T检验方法对目标终端进行粗定位，并进而提出基于T检验的KNN（K-Nearest Neighbour）算法以完成对目标终端的精定位.此方法用有限的样本容量获得较稳定的定位性能分析结果，显著地减少了大量的人力和时间开销.
Abstract：WLAN indoor localization has caught significantly wide attention.In offline phase,the location fingerprint RSS data acquisition often results in blindness and unreliability,and ignores the relations between the required RSS sample capacity and localization performance.To solve this problem,a new T-test based sample capacity optimization approach for WLAN indoor localization is proposed.In offline phase,the Operating Characteristics (OC) function is used to optimize the allowable minimum RSS sample capacity for the fingerprint database construction.In online phase,we perform the rough localization by using the T-test approach,and then propose the T-test based KNN algorithm for the fine localization of target terminal.This method uses a limited sample capacity to obtain a more stable localization performance analysis results,significantly reducing the amount of manpower and time overhead.
[1] SHIN H,CHON Y,KIM Y,et al.MRI:Model-based radio interpolation for indoor war-walking[J].IEEE Transactions on Mobile Computing,2015,14(6):1231-1244.
[2] JUNG S,LEE C,HAN D.Wi-Fi fingerprint-based approaches following Log-distance path loss model for indoor positioning[A].2011 IEEE MTT-S International Microwave Workshop Series on Intelligent Radio for Future Personal Terminals[C].United States:IEEE Computer Society,2011.1-2
[3] 刘剑,龚志恒,吴成东,等.一种基于改进高斯过程隐变量模型的多角度人脸识别算法[J].电子与信息学报,2013,35(9):2033-2039. LIU J,GONG Z H,WU C D,et al.A multi-angle face recognition algorithm based on modified gaussian process latent variable model[J].Journal of Electronics and Information Technology,2013,35(9):2033-2039.(in Chinese)
[4] GU Z,CHEN Z Q,ZHANG Y X,et al.Reducing fingerprint collection for indoor localization[J].Computer Communications,2015,83:56-63.
[5] HAN S,ZHAO C,MENG W X.Cosine similarity based fingerprinting algorithm in WLAN indoor positioning against device diversity[A].IEEE International Conference on Communications,ICC 2015[C].London:Institute of Electrical and Electronics Engineers Inc,2015.2710-2714.
[6] TORRES-SOSPEDRA J,MONTOLIU R,MARTINEZ-USO A,et al.UJⅡndoorLoc:A new multi-building and multi-floor database for WLAN fingerprint-based indoor localization problems[A].5th International Conference on Indoor Positioning and Indoor Navigation,IPIN 2014[C].Korea:Institute of Electrical and Electronics Engineers Inc,2014.261-270.
[7] KARIM T,REMI C,FLORENT R.Fully automatic detection of anomalies on wheels surface using an adaptive accurate model and hypothesis testing theory[A].24th European Signal Processing Conference,EUSIPCO 2016[C].Hungary:European Signal Processing Conference,2016.508-512.
[8] 甘伦知.假设检验中控制第二类错误的探讨[J].统计与决策,2011,346(22):35-37. GAN L Z.On the control of errors of the second kind in hypothesis test[J].Statistics and Decision,2011,346(22):35-37.(in Chinese)
2018, Vol. 46 
