基于Kmeans和动态WKNN的两层Wi-Fi改进定位方法

Yatao Wang; Xinheng Wang; Yuning Dong; Xiaolong Xu

doi:10.14132/j.cnki.1673-5439.2017.05.008

基于Kmeans和动态WKNN的两层Wi-Fi改进定位方法

Translated title of the contribution: Optimized two-layer Wi-Fi location method based on Kmeans and dynamic WKNN

Yatao Wang, Xinheng Wang, Yuning Dong^*, Xiaolong Xu

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

1 Citation (Scopus)

Abstract

The algorithm based on K nearest neighbor (KNN) is widely used in Wi-Fi indoor localization,however traditional KNN algorithm consumes longer pattern matching time,and the use of the fixed K value to search for the nearest neighbor can cause larger positioning errors.To solve these two problems,this paper proposes a new two-layer localization algorithm based on Kmeans clustering and dynamic weighted KNN (WKNN).Experiments show that the algorithm can improve the positioning performance than the traditional KNN,WKNN and enhanced WKNN (EWKNN) algorithms in terms of the computation power and the mean positioning error.Meanwhile,aiming at the temporary 'disappearance' phenomenon in the construction phase of the fingerprinting database,a nearest neighbor interpolation method for the default value is proposed.Experimental

Translated title of the contribution	Optimized two-layer Wi-Fi location method based on Kmeans and dynamic WKNN
Original language	Chinese (Traditional)
Pages (from-to)	41-47
Number of pages	7
Journal	Nanjing Youdian Daxue Xuebao (Ziran Kexue Ban)/Journal of Nanjing University of Posts and Telecommunications (Natural Science)
Volume	37
Issue number	5
DOIs	https://doi.org/10.14132/j.cnki.1673-5439.2017.05.008
Publication status	Published - 1 Oct 2017
Externally published	Yes

Keywords

Enhanced weighted K nearest neighbor(EWKNN)
Kmeans clustering
Wi-Fi indoor positioning

Access to Document

10.14132/j.cnki.1673-5439.2017.05.008

Cite this

@article{c24501db298e42408cddfb9b5e8a2834,

title = "基于Kmeans和动态WKNN的两层Wi-Fi改进定位方法",

abstract = "The algorithm based on K nearest neighbor (KNN) is widely used in Wi-Fi indoor localization,however traditional KNN algorithm consumes longer pattern matching time,and the use of the fixed K value to search for the nearest neighbor can cause larger positioning errors.To solve these two problems,this paper proposes a new two-layer localization algorithm based on Kmeans clustering and dynamic weighted KNN (WKNN).Experiments show that the algorithm can improve the positioning performance than the traditional KNN,WKNN and enhanced WKNN (EWKNN) algorithms in terms of the computation power and the mean positioning error.Meanwhile,aiming at the temporary 'disappearance' phenomenon in the construction phase of the fingerprinting database,a nearest neighbor interpolation method for the default value is proposed.Experimental",

keywords = "Enhanced weighted K nearest neighbor(EWKNN), Kmeans clustering, Wi-Fi indoor positioning",

author = "Yatao Wang and Xinheng Wang and Yuning Dong and Xiaolong Xu",

year = "2017",

month = oct,

day = "1",

doi = "10.14132/j.cnki.1673-5439.2017.05.008",

language = "繁体中文",

volume = "37",

pages = "41--47",

journal = "Nanjing Youdian Daxue Xuebao (Ziran Kexue Ban)/Journal of Nanjing University of Posts and Telecommunications (Natural Science)",

issn = "1673-5439",

number = "5",

}

TY - JOUR

T1 - 基于Kmeans和动态WKNN的两层Wi-Fi改进定位方法

AU - Wang, Yatao

AU - Wang, Xinheng

AU - Dong, Yuning

AU - Xu, Xiaolong

PY - 2017/10/1

Y1 - 2017/10/1

N2 - The algorithm based on K nearest neighbor (KNN) is widely used in Wi-Fi indoor localization,however traditional KNN algorithm consumes longer pattern matching time,and the use of the fixed K value to search for the nearest neighbor can cause larger positioning errors.To solve these two problems,this paper proposes a new two-layer localization algorithm based on Kmeans clustering and dynamic weighted KNN (WKNN).Experiments show that the algorithm can improve the positioning performance than the traditional KNN,WKNN and enhanced WKNN (EWKNN) algorithms in terms of the computation power and the mean positioning error.Meanwhile,aiming at the temporary 'disappearance' phenomenon in the construction phase of the fingerprinting database,a nearest neighbor interpolation method for the default value is proposed.Experimental

AB - The algorithm based on K nearest neighbor (KNN) is widely used in Wi-Fi indoor localization,however traditional KNN algorithm consumes longer pattern matching time,and the use of the fixed K value to search for the nearest neighbor can cause larger positioning errors.To solve these two problems,this paper proposes a new two-layer localization algorithm based on Kmeans clustering and dynamic weighted KNN (WKNN).Experiments show that the algorithm can improve the positioning performance than the traditional KNN,WKNN and enhanced WKNN (EWKNN) algorithms in terms of the computation power and the mean positioning error.Meanwhile,aiming at the temporary 'disappearance' phenomenon in the construction phase of the fingerprinting database,a nearest neighbor interpolation method for the default value is proposed.Experimental

KW - Enhanced weighted K nearest neighbor(EWKNN)

KW - Kmeans clustering

KW - Wi-Fi indoor positioning

UR - http://www.scopus.com/inward/record.url?scp=85055874758&partnerID=8YFLogxK

U2 - 10.14132/j.cnki.1673-5439.2017.05.008

DO - 10.14132/j.cnki.1673-5439.2017.05.008

M3 - 文章

AN - SCOPUS:85055874758

SN - 1673-5439

VL - 37

SP - 41

EP - 47

JO - Nanjing Youdian Daxue Xuebao (Ziran Kexue Ban)/Journal of Nanjing University of Posts and Telecommunications (Natural Science)

JF - Nanjing Youdian Daxue Xuebao (Ziran Kexue Ban)/Journal of Nanjing University of Posts and Telecommunications (Natural Science)

IS - 5

ER -

基于Kmeans和动态WKNN的两层Wi-Fi改进定位方法

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this