基于KPCA和改进GBRT的室内定位算法
|
摘要
针对无线局域网络易受室内环境等因素影响,导致室内定位误差较大的问题,提出了一种基于KPCA(Kernel Principal Component Analysis)和改进GBRT(Gradient Boosting Regression Tree)的室内定位算法。离线训练阶段,通过KPCA算法,利用其核函数提取原始位置指纹向量主成分,最大限度地保留其非线性特征信息,并通过改进GBRT算法,利用其自助抽样法将训练集均匀抽样为多个子训练集,每一子训练集依据GBRT算法构建一个室内定位回归模型,最终根据多个子训练集构成强回归模型;在线阶段,对实时测量的信号强度进行KPCA变换,并根据离线回归模型计算结果的众数预测实时位置。实验结果表明,平均定位误差可低至1.16m,与SMN-PCA和RVM-PLS算法相比,定位准确率分别提高了12.8%和10.1%。
Aiming at the problem that the wireless local area networks is susceptible to some factors like indoor environment,resulting in large indoor positioning errors,an indoor positioning algorithm based on KPCA( Kernel Principal Component Analysis) and improved GBRT( Gradient Boosting Regression Tree) is proposed. In the offline training phase,the KPCA algorithm is used to extract principal components of the original position fingerprint vector by its kernel function,which retain the nonlinear feature information to the maximum extent,and combined with the improved GBRT algorithm,the self-sampling method is used to uniformly sample the training set into multiple subtraining sets,each sub-training set constructs an indoor positioning regression model based on the GBRT algorithm,and finally forms a strong regression model according to multiple sub-training sets; in the online training phase,KPCA transform is performed on the signal strength of the real-time measurement,and the real-time position is predicted based on the mode of results calculated by the offline regression model. The experimental results show that the average positioning error can be as low as 1.16 m,and compared with the SMN-PCA and RVM-PLS algorithm,the positioning accuracy is improved by 12.8% and 10.1%,respectively.
Aiming at the problem that the wireless local area networks is susceptible to some factors like indoor environment,resulting in large indoor positioning errors,an indoor positioning algorithm based on KPCA( Kernel Principal Component Analysis) and improved GBRT( Gradient Boosting Regression Tree) is proposed. In the offline training phase,the KPCA algorithm is used to extract principal components of the original position fingerprint vector by its kernel function,which retain the nonlinear feature information to the maximum extent,and combined with the improved GBRT algorithm,the self-sampling method is used to uniformly sample the training set into multiple subtraining sets,each sub-training set constructs an indoor positioning regression model based on the GBRT algorithm,and finally forms a strong regression model according to multiple sub-training sets; in the online training phase,KPCA transform is performed on the signal strength of the real-time measurement,and the real-time position is predicted based on the mode of results calculated by the offline regression model. The experimental results show that the average positioning error can be as low as 1.16 m,and compared with the SMN-PCA and RVM-PLS algorithm,the positioning accuracy is improved by 12.8% and 10.1%,respectively.
引文
[1] 陈丽娜. WLAN位置指纹室内定位关键技术研究[D]. 上海:华东师范大学,2014.
[2] Xue Weixing,Qiu Weining,Hua Xianghong,et al. Improved Wi-Fi RSSI Measurement for Indoor Localization[J]. IEEE Sensors Journal,2017,17(7):2224-2230.
[3] 陈兵,杨小玲. 一种基于概率密度的WLAN接入点定位的算法[J]. 电子与信息学报,2015,37(4):855-862.
[4] 王杨,赵红东. 室内定位技术综述及发展前景展望[J]. 测控技术,2016,35(7):1-3,8.
[5] 孟祥武,胡勋,王立才,等. 移动推荐系统及其应用[J]. 软件学报,2013,24(1):91-108.
[6] 王涵,徐凌伟,肖平平,等. 移动无线传感器网络系统的物理层安全性能分析[J]. 传感技术学报,2018,31(7):1108-1112.
[7] 徐玉滨,邓志安,马琳. 基于核直接判别分析和支持向量回归的WLAN室内定位算法[J]. 电子与信息学报,2011,33(4):896-901.
[8] Shih-Hau Fang,Tsung-Nan Lin. Principal Component Localization in Indoor WLAN Environments[J]. IEEE Transactions on Mobile Computing,2012,11(1):100-110.
[9] 张勇,黄杰,徐科宇. 基于PCA-LSSVR算法的WLAN室内定位方法[J]. 仪器仪表学报,2015,36(2):408-414.
[10] Basiouny Y,Arafa M,Amany M Sarhan. Enhancing Wi-Fi Fingerprinting for Indoor Positioning System Using Single Multiplicative Neuron and PCA Algorithm[C]//2017 12th International Conference Computer Engineering and Systems,Cairo:IEEE,2017:295-305.
[11] Lu Xiaoqing,Liu Haitao,Liu Feng. A Novel Algorithm for Enhancing Accuracy of Indoor Position Estimation[C]//Proceeding of the 11th World Congress on Intelligent Control and Automation,Shenyang:IEEE,2014:5528-5533.
[12] 周红亮,周先存,陈孟元. 基于稀疏表示和位置相关性的室内定位算法[J]. 传感技术学报,2018,31(2):265-270.
[13] Chen Chen,Wang Yujie,Zhang Yong,et al. Indoor Positioning Algorithm Based on Nonlinear PLS Integrated with RVM[J]. IEEE Sensors Journal,2018,18(2):660-668.
[14] 焦秀珍,石子烨. 基于格拉布斯校验法的陆军模拟训练学员成绩评估方法[J]. 中国电子科学研究院学报,2016,11(3):283-287.
[15] 付华,王馨蕊,王志军,等. 基于PCA和PSO-ELM的煤与瓦斯突出软测量研究[J]. 传感技术学报,2014,27(12):1710-1715.
[16] Elkhadir Z,Chougdali K,Benattou M. Intrusion Detection System Using PCA and Kernel PCA Methods[M]. Springer International Publishing,2016:13-22.
[17] 孙克雷,邓仙荣. 一种改进的基于梯度提升回归算法的O2O电子商务推荐模型[J]. 安徽建筑大学学报,2016,24(2):87-91.
[18] 王守相,刘天宇. 计及用电模式的居民负荷梯度提升树分类识别方法[J]. 电力系统及其自动化学报,2017,29(9):27-33.李新春(1963-),男,辽宁喀左人,高级工程师,硕士生导师,主要研究方向为室内无线定位技术,图像处理与模式识别等,550128966@qq.com;房梽斅(1993-),男,辽宁阜新人,硕士研究生,主要研究方向为室内无线定位技术,1914023617@qq.com; 张春华(1994-),女,辽宁朝阳人,硕士研究生,主要研究方向:图像处理与模式识别,1226617885@qq.com。
[2] Xue Weixing,Qiu Weining,Hua Xianghong,et al. Improved Wi-Fi RSSI Measurement for Indoor Localization[J]. IEEE Sensors Journal,2017,17(7):2224-2230.
[3] 陈兵,杨小玲. 一种基于概率密度的WLAN接入点定位的算法[J]. 电子与信息学报,2015,37(4):855-862.
[4] 王杨,赵红东. 室内定位技术综述及发展前景展望[J]. 测控技术,2016,35(7):1-3,8.
[5] 孟祥武,胡勋,王立才,等. 移动推荐系统及其应用[J]. 软件学报,2013,24(1):91-108.
[6] 王涵,徐凌伟,肖平平,等. 移动无线传感器网络系统的物理层安全性能分析[J]. 传感技术学报,2018,31(7):1108-1112.
[7] 徐玉滨,邓志安,马琳. 基于核直接判别分析和支持向量回归的WLAN室内定位算法[J]. 电子与信息学报,2011,33(4):896-901.
[8] Shih-Hau Fang,Tsung-Nan Lin. Principal Component Localization in Indoor WLAN Environments[J]. IEEE Transactions on Mobile Computing,2012,11(1):100-110.
[9] 张勇,黄杰,徐科宇. 基于PCA-LSSVR算法的WLAN室内定位方法[J]. 仪器仪表学报,2015,36(2):408-414.
[10] Basiouny Y,Arafa M,Amany M Sarhan. Enhancing Wi-Fi Fingerprinting for Indoor Positioning System Using Single Multiplicative Neuron and PCA Algorithm[C]//2017 12th International Conference Computer Engineering and Systems,Cairo:IEEE,2017:295-305.
[11] Lu Xiaoqing,Liu Haitao,Liu Feng. A Novel Algorithm for Enhancing Accuracy of Indoor Position Estimation[C]//Proceeding of the 11th World Congress on Intelligent Control and Automation,Shenyang:IEEE,2014:5528-5533.
[12] 周红亮,周先存,陈孟元. 基于稀疏表示和位置相关性的室内定位算法[J]. 传感技术学报,2018,31(2):265-270.
[13] Chen Chen,Wang Yujie,Zhang Yong,et al. Indoor Positioning Algorithm Based on Nonlinear PLS Integrated with RVM[J]. IEEE Sensors Journal,2018,18(2):660-668.
[14] 焦秀珍,石子烨. 基于格拉布斯校验法的陆军模拟训练学员成绩评估方法[J]. 中国电子科学研究院学报,2016,11(3):283-287.
[15] 付华,王馨蕊,王志军,等. 基于PCA和PSO-ELM的煤与瓦斯突出软测量研究[J]. 传感技术学报,2014,27(12):1710-1715.
[16] Elkhadir Z,Chougdali K,Benattou M. Intrusion Detection System Using PCA and Kernel PCA Methods[M]. Springer International Publishing,2016:13-22.
[17] 孙克雷,邓仙荣. 一种改进的基于梯度提升回归算法的O2O电子商务推荐模型[J]. 安徽建筑大学学报,2016,24(2):87-91.
[18] 王守相,刘天宇. 计及用电模式的居民负荷梯度提升树分类识别方法[J]. 电力系统及其自动化学报,2017,29(9):27-33.李新春(1963-),男,辽宁喀左人,高级工程师,硕士生导师,主要研究方向为室内无线定位技术,图像处理与模式识别等,550128966@qq.com;房梽斅(1993-),男,辽宁阜新人,硕士研究生,主要研究方向为室内无线定位技术,1914023617@qq.com; 张春华(1994-),女,辽宁朝阳人,硕士研究生,主要研究方向:图像处理与模式识别,1226617885@qq.com。