结合空间划分和支持向量机的两级定位算法
|
摘要
目前室内定位的主流方法是根据WiFi指纹模式匹配来确定目标位置,但由于室内环境的复杂性和WiFi信号的不稳定性,其性能尚不能满足高精度室内定位的需求.为提高定位精度,提出一种基于空间划分和支持向量机(Support Vector Ma-chines,SVM)的两级室内定位算法.该算法首先采用优化K-means聚类算法,对定位区域的WiFi指纹进行聚类,并据此进行空间划分产生子区域;然后采用SVM实现两级WiFi指纹定位:第一级采用SVM分类确定目标所在子区域,第二级在子区域内采用SVM回归确定目标精确位置坐标.实验表明,提出的基于空间划分的两级定位精度明显优于单级定位,提出的空间划分算法优于当前其它空间划分算法.
WiFi fingerprinting is a mainstream indoor localization method. Due to complexity of indoor environments and instability of WiFi signals,performance of WiFi fingerprinting cannot meet the requirement of high accuracy. In order to improve the accuracy,a two-layer WiFi fingerprinting based on space partitioning and support vector machines( SVM) is proposed. The algorithm first applies an optimized K-means clustering algorithm to partition the localization space to subregions according to WiFi fingerprints,then applies SVMto achieve two-layer WiFi fingerprinting. The first layer determines the subregion through SVMclassification and the second layer determines the exact location coordinate through SVMregression in the determined subregion. Evaluations showthat the proposed two-layer WiFi fingerprinting based on space partitioning outperforms one-layer WiFi fingerprinting,and the proposed space partitioning method outperforms other existing space partitioning methods.
WiFi fingerprinting is a mainstream indoor localization method. Due to complexity of indoor environments and instability of WiFi signals,performance of WiFi fingerprinting cannot meet the requirement of high accuracy. In order to improve the accuracy,a two-layer WiFi fingerprinting based on space partitioning and support vector machines( SVM) is proposed. The algorithm first applies an optimized K-means clustering algorithm to partition the localization space to subregions according to WiFi fingerprints,then applies SVMto achieve two-layer WiFi fingerprinting. The first layer determines the subregion through SVMclassification and the second layer determines the exact location coordinate through SVMregression in the determined subregion. Evaluations showthat the proposed two-layer WiFi fingerprinting based on space partitioning outperforms one-layer WiFi fingerprinting,and the proposed space partitioning method outperforms other existing space partitioning methods.
引文
[1]Zhang Xiao-jun,Liu Wen-yuan,Liu Yong-shan,et al.An indoor location method based on speech recognition[J].Journal of Chinese Computer Systems,2016,37(8):1883-1888.
[2]Zhao Chu-han,Zhang Hong-ming,Song Jian.Fingerprint based indoor visible light positioning method[J/OL].Chinese Journal of Lasers,2018,45(8):1-14.
[3]Hu Xu-ke,Shang Jian-ga,Gu Fu-qiang,et al.Development of indoor and outdoor seamless positioning prototype system integrating GPS and WiFi[J].Journal of Chinese Computer Systems,2014,35(2):428-432.
[4]Borenovi M N,Neskovic A M.Positioning in WLAN environment by use of artificial neural netw orks and space partitioning[J].Annals of Telecommunications Annales des TéléCommunications,2009,64(9-10):665-676.
[5]Vahidnia M H,Malek M R,Mohammadi N,et al.A hierarchical signal-space partitioning technique for indoor positioning w ith WLAN to support location-aw areness in mobile map services[J].Wireless Personal Communications,2013,69(2):689-719.
[6]Lee C W,Lin T N,Fang S H,et al.A novel clustering-based approach of indoor location fingerprinting[C].Personal Indoor and M obile Radio Communications(PIM RC),IEEE 24th International Symposium on,2013:3191-3196.
[7]Mo Y,Zhang Z,Meng W,et al.A spatial division clustering method and low dimensional feature extraction technique based indoor positioning system[J].Sensors,2014,14(1):1850-1876.
[8]Vapnik V.The nature of statistical learning theory[M].Springer Science&Business M edia,2013.
[9]Friedman J,Hastie T,Tibshirani R.The elements of statistical learning[M].New York:Springer Series in Statistics,2001.
[10]Hsu C W,Lin C J.A comparison of methods for multiclass support vector machines[J].IEEE Transactions on Neural Netw orks,2002,13(2):415-425.
[11]Scholkopf B,Smola A J,Williamson R C,et al.New support vector algorithms[J].Neural Computation,2000,12(5):1207-1245.
[1]张晓军,刘文远,刘永山,等.一种基于语音识别的室内定位方法[J].小型微型计算机系统,2016,37(8):1883-1888.
[2]赵楚韩,张洪明,宋健.基于指纹的室内可见光定位方法[J/OL].中国激光,2018,45(8):1-14.
[3]胡旭科,尚建嘎,古富强,等.融合GPS与Wi-Fi的室内外无缝定位原型系统研制[J].小型微型计算机系统,2014,35(2):428-432.
[2]Zhao Chu-han,Zhang Hong-ming,Song Jian.Fingerprint based indoor visible light positioning method[J/OL].Chinese Journal of Lasers,2018,45(8):1-14.
[3]Hu Xu-ke,Shang Jian-ga,Gu Fu-qiang,et al.Development of indoor and outdoor seamless positioning prototype system integrating GPS and WiFi[J].Journal of Chinese Computer Systems,2014,35(2):428-432.
[4]Borenovi M N,Neskovic A M.Positioning in WLAN environment by use of artificial neural netw orks and space partitioning[J].Annals of Telecommunications Annales des TéléCommunications,2009,64(9-10):665-676.
[5]Vahidnia M H,Malek M R,Mohammadi N,et al.A hierarchical signal-space partitioning technique for indoor positioning w ith WLAN to support location-aw areness in mobile map services[J].Wireless Personal Communications,2013,69(2):689-719.
[6]Lee C W,Lin T N,Fang S H,et al.A novel clustering-based approach of indoor location fingerprinting[C].Personal Indoor and M obile Radio Communications(PIM RC),IEEE 24th International Symposium on,2013:3191-3196.
[7]Mo Y,Zhang Z,Meng W,et al.A spatial division clustering method and low dimensional feature extraction technique based indoor positioning system[J].Sensors,2014,14(1):1850-1876.
[8]Vapnik V.The nature of statistical learning theory[M].Springer Science&Business M edia,2013.
[9]Friedman J,Hastie T,Tibshirani R.The elements of statistical learning[M].New York:Springer Series in Statistics,2001.
[10]Hsu C W,Lin C J.A comparison of methods for multiclass support vector machines[J].IEEE Transactions on Neural Netw orks,2002,13(2):415-425.
[11]Scholkopf B,Smola A J,Williamson R C,et al.New support vector algorithms[J].Neural Computation,2000,12(5):1207-1245.
[1]张晓军,刘文远,刘永山,等.一种基于语音识别的室内定位方法[J].小型微型计算机系统,2016,37(8):1883-1888.
[2]赵楚韩,张洪明,宋健.基于指纹的室内可见光定位方法[J/OL].中国激光,2018,45(8):1-14.
[3]胡旭科,尚建嘎,古富强,等.融合GPS与Wi-Fi的室内外无缝定位原型系统研制[J].小型微型计算机系统,2014,35(2):428-432.