CSI indoor positioning algorithm based on GMM-DBC
-
摘要: 针对贝叶斯室内定位技术存在定位精度低及时间复杂度较高的问题，提出了一种基于高斯混合模型和密度聚类(GMM-DBC)的信道状态信息(CSI)定位算法. 通过对分模型参数的初次估计构建GMM概率分布模型并进行误差计算；引入确定分模型个数(DSM)策略，结合误差计算结果更新GMM模型参数，减小由模型精度引起的定位误差；基于不同参考点的分布特征，判断参考点间紧密程度，将紧密相连的参考点划为一类，减小搜索范围，降低时间复杂度；根据分簇结果，利用改进的贝叶斯概率算法进行权值计算，得到最终定位结果. 实验结果表明：所提算法能较好地提高定位精度，降低时间复杂度.
-
关键词:
- 室内定位 /
- 信道状态信息(CSI) /
- 贝叶斯概率算法 /
- 高斯混合模型 /
- 密度聚类(DBC)
Abstract: A channel state information (CSI) location algorithm based on the Gaussian mixture model and density based clustering (GMM-DBC) is proposed to deal with the low positioning accuracy and high time complexity of Bayesian indoor positioning technology. The GMM probability distribution model is constructed through the initial estimation of the parameters of the sub-model, and the errors were calculated. Then, introduce a strategy to determine the number of sub-models (DSM) to update the parameters of the GMM and reduce the localization error caused by the model accuracy. The tightness between reference points is judged on the basis of the distribution characteristics of different reference points, and the closely connected reference points are classified into one class to reduce the search scope and time complexity. The weights are calculated via the improved Bayesian probability algorithm according to the clustering results, so as to obtain the final positioning results. The experimental results show that the proposed algorithm can well improve the positioning accuracy and reduce the time complexity. -
表 1
$k$ 值对累积误差的影响序号 $k$值 累积误差 1 $k$=2 15.836 $k$=3 15.854 $k$=4 15.860 $k$=5 15.881 2 $k$=2 11.620 $k$=3 11.566 $k$=4 11.568 $k$=5 11.567 3 $k$=2 15.855 $k$=3 15.845 $k$=4 15.838 $k$=5 15.840 
下载: 导出CSV
表 2 不同方法定位精度对比
算法 最大定位
误差/m最小定位
误差/m平均定位
误差/m定位时间/
s本文算法 4.052 6 0.097 2 1.012 3 13.2 CNN 4.859 2 0.218 4 1.268 6 15.7 WKNN-
贝叶斯6.772 8 0.225 3 1.667 8 13.5 传统贝叶斯 5.942 1 0.463 4 1.779 7 17.6
下载: 导出CSV
-
[1] ZHOU T Y, LIAN B W, YI Z, et al. Amp-Phi: a CSI-based indoor positioning system[J]. International journal of pattern recognition and artificial intelligence, 2018, 32(21): 1-22. DOI: 10.1142/S0218001418580053 [2] KIM G T, ZAR A K, AUNG M S, et al. Review of indoor positioning: radio wave technology[J]. Applied sciences, 2020, 11(1): 279. DOI: 10.3390/app11010279 [3] YANG C C, SHAO H R. WiFi-based indoor positioning[J]. IEEE communications magazine, 2015, 53(3): 150-157. DOI: 10.1109/MCOM.2015.7060497 [4] BOORANAWONG A, SENGCHUAI K, BURANAPANICHKIT D, et al. RSS-based indoor localization using multi-lateration with zone selection and virtual position-based compensation methods[J]. IEEE access, 2021(99): 1-1. DOI: 10.1109/ACCESS.2021.3068295 [5] ZHANG J, MAO H Q. WKNN indoor positioning method based on spatial feature partition and basketball motion capture[J]. Alexandria engineering journal, 2022, 61(1): 125-134. DOI: 10.1016/j.aej.2021.04.078 [6] 吴哲夫, 徐强, 王中友, 等. 基于信道状态信息的无源室内定位[J]. 哈尔滨工程大学学报, 2017, 38(8): 1328-1334. [7] 聂大惟, 朱海, 吴飞, 等. 基于RSSI概率分布与贝叶斯估计的加权定位方法[J]. 全球定位系统, 2022, 47(02): 52-59. [8] 刘帅, 王旭东, 吴楠. 一种基于卷积神经网络的CSI指纹室内定位方法[J]. 工程科学学报, 2021, 43(11): 1512-1521. [9] 杨如民, 陈敏, 余成波. 基于贝叶斯概率优化的Wi-Fi室内定位算法[J]. 计算机应用与软件, 2021, 38(2): 97-144. DOI: 10.3969/j.issn.1000-386x.2021.02.017 [10] TSE D N C, VISWANATH P. Fundamentals of wireless communication (hardcover)[M]. Britain: Cambridge University Press, 2005. DOI: 10.1017/CBO9780511807213 [11] ASHERI H, HOSSEINI R, ARAABI B N. A new EM algorithm for flexibly tied GMMs with large number of components[J]. Pattern recognition, 2021(114): 107836. DOI: 10.1016/j.patcog.2021.107836 [12] SUN H J, WANG S R. Measuring the component overlapping in the gaussian mixture model[J]. Data mining and knowledge discovery, 2011, 23(3): 479-502. DOI: 10.1007/s10618-011-0212-3 [13] 柳兴旺. 一种有效估计负二项分布参数的EM算法及其应用[D]. 哈尔滨: 哈尔滨工业大学, 2019. [14] TSENG C H, YEN J S. Enhanced gaussian mixture model for indoor positioning accuracy[C]// International Computer Symposium (ICS), 2016: 462-466. DOI: 10.1109/ICS.2016.0099 [15] KUMAR V, DHOK S B, TRIPATHI R, et al. A review study of hierarchical clustering algorithms for wireless sensor networks[J]. International journal of computer science issues (IJCSI), 2014, 11(3): 92-101. [16] 宋董飞, 徐华. DBSCAN算法研究及并行化实现[J]. 计算机工程与应用, 2018, 54(24): 52-56. DOI: 10.3778/j.issn.1002-8331.1808-0423 [17] BIRANT D, KUT A. ST-DBSCAN: an algorithm for clustering spatial-temporal data[J]. Data and knowledge engineering, 2007, 60(1): 208-221. DOI: 10.1016/j.datak.2006.01.013 [18] 阮曙芬. 属性加权多项式朴素贝叶斯算法及应用研究[D]. 武汉: 中国地质大学, 2021. [19] IBRAHIM M, YOUSSEF M. CellSense: a probabilistic RSS-based GSM positioning system[C]// IEEE Global Telecommunications Conference GLOBECOM, 2010. DOI: 10.1109/GLOCOM.2010.5683779 [20] 李新春, 赵忠婷, 于洪仕. 基于局部线性嵌入和梯度提升决策树的信道状态信息室内指纹定位算法研究[J]. 激光与光电子学进展, 2022, 59(2): 392-402. -
点击查看大图
图(10) / 表(2)
计量
- 文章访问数: 228
- HTML全文浏览量: 168
- PDF下载量: 33
- 被引次数: 0
