您的位置：山东大学 -> 科技期刊社 -> 《山东大学学报（工学版）》

山东大学学报 (工学版) ›› 2018, Vol. 48 ›› Issue (5): 95-102.doi: 10.6040/j.issn.1672-3961.0.2018.169

• 控制科学与工程 • 上一篇    下一篇

基于集成深度神经网络的室内无线定位

沈冬冬(),周风余*(),栗梦媛,王淑倩,郭仁和   

  1. 山东大学控制科学与工程学院, 山东 济南 250061
  • 收稿日期:2018-05-02 出版日期:2018-10-01 发布日期:2018-05-02
  • 通讯作者: 周风余 E-mail:shendongdong724@foxmail.com;zhoufengyu@sdu.edu.cn
  • 作者简介:沈冬冬(1992—)，男，安徽安庆人，硕士研究生，主要研究方向为深度学习. E-mail:shendongdong724@foxmail.com
  • 基金资助:
    国家重点研发计划(2017YFB1302400);国家自然科学基金资助项目(61773242);山东省重大科技创新工程资助项目(2017CXGC0926);山东省重点研发计划资助项目(公益类专项)(2017GGX30133)

Indoor wireless positioning based on ensemble deep neural network

Dongdong SHEN(),Fengyu ZHOU*(),Mengyuan LI,Shuqian WANG,Renhe GUO   

  1. School of Control Science and Engineering, Shandong University, Jinan 250061, Shandong, China
  • Received:2018-05-02 Online:2018-10-01 Published:2018-05-02
  • Contact: Fengyu ZHOU E-mail:shendongdong724@foxmail.com;zhoufengyu@sdu.edu.cn
  • Supported by:
    国家重点研发计划(2017YFB1302400);国家自然科学基金资助项目(61773242);山东省重大科技创新工程资助项目(2017CXGC0926);山东省重点研发计划资助项目(公益类专项)(2017GGX30133)

RichHTML

29

PDF (PC)

367

摘要：

针对传统无线定位模型对指纹数据库容错性低、抗噪能力弱等问题,提出一种基于数据融合的集成深度神经网络无线定位方法,从原始指纹数据库中按照一定比例随机取样生成各基学习器的训练数据,能够有效克服异常样本与有噪数据对无线定位系统带来的干扰;在指纹数据库构建过程中,提出Gauss-Occupied (G-O)数据扩充方法以解决无线指纹数据库样本容量小的局限,大幅度降低人工采集的成本,进一步提高样本空间的表征范围。试验结果表明:提出的模型不仅能够有效提高无线定位系统的平均定位精度与抗噪能力,而且能够明显降低定位过程中出现的单点最大误差。

关键词: 无线指纹定位, 数据集扩充, 人工采集, 深度神经网络, 集成学习

Abstract:

Because of the low fault tolerance and weak anti-noise ability of fingerprint database in traditional wireless positioning model, an ensemble deep neural network wireless positioning method based on data fusion was proposed. This method could effectively overcome the interference caused by abnormal samples and noisy data on the wireless positioning system by sampling from the original fingerprint database randomly to generate train data for each base learner. During the process of fingerprint database construction, the Gauss-Occupied (G-O) data expansion method was proposed to solve the limitation of the small sample size of the wireless fingerprint database and decrease the cost of manual acquisition sharply, which increased the scope of the sample′s characterization. The results of the experiment showed that the proposed ensemble deep neural network wireless positioning model could not only improve the average positioning accuracy and the anti-noise ability of the wireless positioning system, but also reduce the maximum single point error in the positioning process.

Key words: wireless fingerprint positioning, data set expansion, artificial collection, deep neural network, ensemble learning

中图分类号: 

  • TP391

图1

无线信号随时间变化结果"

图2

高斯噪声信号"

图3

数据集扩充流程"

图4

深度神经网络基学习器"

图5

集成深度神经网络训练模型"

图6

某学生宿舍楼二层楼道平面布局图"

表1

G-O数据集扩充方法平均定位误差试验对比"

m
数据集 Error1 Error2
Testset1 0.401 0.513
Testset2 0.672 0.768
Testset3 0.614 0.756
Testset4 0.674 0.848

图7

平均定位误差随基学习器个数变化的结果"

图8

基学习器与集成学习器的平均定位误差对比结果"

表2

单一深度神经网络与集成深度神经网络平均定位误差试验对比"

m
模型 Testset1 Testset2 Testset3 Testset4
GO_DNN_A 0.390 0.652 0.628 0.663
GO_DNN_B 0.403 0.719 0.612 0.717
GO_DNN_C 0.396 0.688 0.601 0.694
GO_DNN_D 0.385 0.669 0.595 0.678
Ensemble 0.329 0.563 0.516 0.552

表3

单一深度神经网络与集成深度神经网络最大定位误差试验对比"

m
模型 Testset1 Testset2 Testset3 Testset4
GO_DNN_A 1.944 2.170 1.912 1.970
GO_DNN_B 1.855 1.906 2.330 2.653
GO_DNN_C 1.799 1.929 2.450 2.412
GO_DNN_D 2.044 1.891 2.184 2.182
Ensemble 1.610 1.654 1.500 1.556

表4

各算法平均定位误差试验对比"

m
模型 Testset1 Testset2 Testset3 Testset4
KNN 1.974 2.648 2.437 2.921
SVM 1.812 2.437 2.159 2.766
Ensemble 0.329 0.563 0.516 0.552

表5

各算法最大定位误差试验对比"

m
模型 Testset1 Testset2 Testset3 Testset4
KNN 5.725 6.177 5.870 7.474
SVM 4.874 4.946 4.324 5.489
Ensemble 1.610 1.654 1.500 1.556
1 TORTEEKA P, XIU C. Indoor positioning based on WiFi fingerprint technique using fuzzy K-nearest neighbor[C]//International Bhurban Conference on Applied Sciences and Technology. Islamabad, Pakistan: IEEE, 2014: 461-465.
2 YANG R, ZHANG H. RSSI-based fingerprint positioning system for indoor wireless network[M]//Intelligent Computing in Smart Grid and Electrical Vehicles. Berlin Heidelberg, Germany:Springer, 2014:313-319.
3 曾碧, 毛勤. 改进的室内三维模糊位置指纹定位算法[J]. 山东大学学报(工学版), 2015, 45 (3): 22- 27.
ZENG Bi , MAO Qin . Improved indoor 3-D fuzzy position fingerprint localization algorithm[J]. Journal of Shandong University(Engineering Science), 2015, 45 (3): 22- 27.
4 ZHANG W, HUA X, YU K, et al. Domain clustering based WiFi indoor positioning algorithm[C]//International Conference on Indoor Positioning and Indoor Navigation. Alcalá de Henares, Spain: IEEE, 2016: 1-5.
5 XIE Y , WANG Y , NALLANATHAN A , et al. An improved K-Nearest-Neighbor indoor localization method based on spearman distance[J]. IEEE Signal Processing Letters, 2016, 23 (3): 351- 355.
doi: 10.1109/LSP.2016.2519607
6 GE X, QU Z. Optimization WIFI indoor positioning KNN algorithm location-based fingerprint[C]//IEEE International Conference on Software Engineering and Service Science.Beijing, China: IEEE, 2017: 135-137.
7 ZHU Y J , DENG Z L , LIU W L , et al. Multi-classification algorithm for indoor positioning based on support vector machine[J]. Computer Science, 2012, 39 (4): 32- 35.
8 LIU S, SI P, XU M, et al. Edge big data-enabled low-cost indoor localization based on Bayesian analysis of RSS[C]//Wireless Communications and Networking Conference (WCNC). San Francisco, USA: IEEE, 2017: 1-6.
9 DING G, TAN Z, ZHANG J, et al. Fingerprinting localization based on affinity propagation clustering and artificial neural networks[C]//Wireless Communications and Networking Conference. Shanghai, China: IEEE, 2013: 2317-2322.
10 刘侃, 张伟, 张伟东, 等. 一种基于深度神经网络的无线定位方法[J]. 计算机工程, 2016, 42 (7): 82- 85.
doi: 10.3969/j.issn.1000-3428.2016.07.014
LIU Kan , ZHANG Wei , ZHANG Weidong , et al. A wireless positioning method based on deep neural network[J]. Computer Engineering, 2016, 42 (7): 82- 85.
doi: 10.3969/j.issn.1000-3428.2016.07.014
11 QIU X, ZHANG L, REN Y, et al. Ensemble deep learning for regression and time series forecasting[C]//Computational Intelligence in Ensemble Learning. Orlando, USA: IEEE, 2015: 1-6.
12 ZHOU X, XIE L, ZHANG P, et al. An ensemble of deep neural networks for object tracking[C]//IEEE International Conference on Image Processing. Paris, France: IEEE, 2015: 843-847.
13 TSENG P H , FENG K T , LIN Y C , et al. Wireless location tracking algorithms for environments with insufficient signal sources[J]. IEEE Transactions on Mobile Computing, 2009, 8 (12): 1676- 1689.
doi: 10.1109/TMC.2009.75
14 WANG G , SO M C , LI Y . Robust convex approximation methods for TDOA-based localization under NLOS conditions[J]. IEEE Transactions on Signal Processing, 2016, 64 (13): 3281- 3296.
doi: 10.1109/TSP.2016.2539139
15 TOMIC S , BEKO M , RUI D . 3-D target localization in wireless sensor network using RSS and AOA measurements[J]. IEEE Transactions on Vehicular Technology, 2017, 66 (4): 1.
16 马春龙, 张启英. 基于高斯函数及分批估计融合理论的无线网络定位算法[J]. 长春工业大学学报(自然科学版), 2012, 33 (1): 59- 63.
doi: 10.3969/j.issn.1674-1374-B.2012.01.014
MA Chunlong , ZHANG Qiying . Wireless network positioning algorithm based on Gauss function and batch estimation fusion[J]. Journal of Changchun University of Technology(Natural Science Edition), 2012, 33 (1): 59- 63.
doi: 10.3969/j.issn.1674-1374-B.2012.01.014
17 VINCENT P, LAROCHELLE H, BENGIO Y, et al. Extracting and composing robust features with denoising autoencoders[C]//International Conference on Machine Learning. Helsinki, Finland: ACM, 2008: 1096-1103.
18 DAHL G E, SAINATH T N, HINTON G E. Improving deep neural networks for LVCSR using rectified linear units and dropout[C]//IEEE International Conference on Acoustics, Speech and Signal Processing.Vancouver. Canada: IEEE, 2013: 8609-8613.
19 SRIVASTAVA N , HINTON G , KRIZHEVSKY A , et al. Dropout: a simple way to prevent neural networks from overfitting[J]. Journal of Machine Learning Research, 2014, 15 (1): 1929- 1958.
20 REN Y , ZHANG L , SUGANNTHAN P N . Ensemble classification and regression-recent developments, applications and future directions[J]. IEEE Computational Intelligence Magazine, 2016, 11 (1): 41- 53.
doi: 10.1109/MCI.2015.2471235
21 MORETTI F , PIZZUTI S , PANZIERI S , et al. Urban traffic flow forecasting through statistical and neural network bagging ensemble hybrid modeling[J]. Neurocomputing, 2015, 167 (C): 3- 7.
22 王田苗, 陶永, 魏洪兴, 等. 基于智能空间的家庭服务机器人混合定位方法[J]. 北京航空航天大学学报, 2009, 35 (2): 231- 235.
WANG Tianmiao , TAO Yong , WEI Hongxing , et al. Hybrid location method for home service robot based on intelligent space[J]. Journal of Beijing University of Aeronautics and Astronautics, 2009, 35 (2): 231- 235.
[1] 张璞,刘畅,王永. 基于特征融合和集成学习的建议语句分类模型[J]. 山东大学学报 (工学版), 2018, 48(5): 47-54.
[2] 唐乐爽,田国会,黄彬. 一种基于DSmT推理的物品融合识别算法[J]. 山东大学学报(工学版), 2018, 48(1): 50-56.
[3] 刘帆,陈泽华,柴晶. 一种基于深度神经网络模型的多聚焦图像融合方法[J]. 山东大学学报(工学版), 2016, 46(3): 7-13.
[4] 王立宏,李强. 旅行商问题的一种选择性集成求解方法[J]. 山东大学学报(工学版), 2016, 46(1): 42-48.
[5] 陈大伟,闫昭*,刘昊岩. SVD系列算法在评分预测中的过拟合现象[J]. 山东大学学报(工学版), 2014, 44(3): 15-21.
[6] 房晓南1,2,张化祥1,2*,高爽1,2. 基于SMOTE和随机森林的Web spam检测[J]. 山东大学学报(工学版), 2013, 43(1): 22-27.
[7] 张伶卫,万文强. 基于云计算平台的代价敏感集成学习算法研究[J]. 山东大学学报(工学版), 2012, 42(4): 19-23.
[8] 谢伙生,刘敏. 一种基于主动学习的集成协同训练算法[J]. 山东大学学报(工学版), 2012, 42(3): 1-5.
[9] 李小斌1, 李世银2. 时间序列早期分类的多分类器集成方法[J]. 山东大学学报(工学版), 2011, 41(4): 73-78.
[10] 李霞1,王连喜2,蒋盛益1. 面向不平衡问题的集成特征选择[J]. 山东大学学报(工学版), 2011, 41(3): 7-11.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
[1] 何东之, 张吉沣, 赵鹏飞. 不确定性传播算法的MapReduce并行化实现[J]. 山东大学学报(工学版), 0, (): 22 -28 .
[2] 黄劲潮. 基于快速区域建议网络的图像多目标分割算法[J]. 山东大学学报(工学版), 2018, 48(4): 20 -26 .
[3] 唐庆顺,金璐,李国栋,吴春富. 基于自适应终端滑模控制器的机械手跟踪控制[J]. 山东大学学报(工学版), 2016, 46(5): 45 -53 .
[4] 戴平,李宁*. 一种基于SVM的快速特征选择方法[J]. 山东大学学报(工学版), 2010, 40(5): 60 -65 .
[5] 张建明, 刘泉声, 唐志成, 占婷, 蒋亚龙. 考虑剪切变形历史影响的节理峰值剪切强度准则[J]. 山东大学学报(工学版), 0, (): 77 -81 .
[6] 王换,周忠眉. 一种基于聚类的过抽样算法[J]. 山东大学学报(工学版), 2018, 48(3): 134 -139 .
[7] 阳爱民1,周咏梅1,邓河2,周剑峰3. 一种网络流量分类特征的产生及选择方法[J]. 山东大学学报(工学版), 2010, 40(5): 1 -7 .
[8] 尤鸣宇,陈燕,李国正. 不均衡问题中的特征选择新算法:Im-IG[J]. 山东大学学报(工学版), 2010, 40(5): 123 -128 .
[9] 吴国瑶,马立勇. 基于B样条FFD模型配准的虹膜图像融合方法[J]. 山东大学学报(工学版), 2010, 40(5): 24 -27 .
[10] 肖乔, 裴继红, 王荔霞, 龚志成. 基于多通道Gabor滤波模糊融合的遥感图像舰船检测[J]. 山东大学学报(工学版), 0, (): 29 -35 .
版权所有 © 2018 《山东大学学报(工学版)》编辑部　
地址: 济南市山大南路27号（250100）　 电话: 0531-88366735　 E-mail: xbgxb@sdu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发