基于雾计算的物联网高维度数据压缩算法  

Research on high-dimensional data compression method of Internet of Things based on observation matrix optimization

作　　者：袁媛[1] 吕建新 杜传祥[1] 魏秀岭[1] YUAN Yuan;LYU Jianxin;DU Chuanxiang;WEI Xiuling(School of Engineering and Technology,Xi’an Siyuan College,Xi’an 710038,China)

机构地区：[1]西安思源学院工学院,陕西西安710038

出　　处：《西安邮电大学学报》2022年第3期39-45,共7页Journal of Xi’an University of Posts and Telecommunications

基　　金：陕西省自然科学基础研究计划项目(21JK0854)。

摘　　要：为了降低云-雾节点之间的通信消耗,对物联网数据融合计算效率问题进行研究,提出一种基于雾计算的物联网高维度数据压缩算法。在空时压缩机制基础上,建立了雾计算的离线与实时混合数据分析架构,充分考虑高维度数据的空时相关性和观测矩阵的自相关性。通过构造克罗内克(Kronecker)联合观测矩阵,并利用最优化理论对其进行求解,获得该观测矩阵最优值。最后,基于Intel Lab Data公开数据集对提出的算法进行实验验证。实验结果表明,相比于有效数据聚合(Effective Data Aggregating Method based on Compressive Sensing,EDAM)算法和基于动态环的路由方案实现寿命最大化(Lifetime Maximization Dynamic Ring-based Routing Scheme,LMDRS)算法,基于雾计算的观测矩阵优化压缩算法收敛性较好、压缩误差低,在相同的压缩率情况下,所提算法的相对重构误差最小。In order to reduce the communication consumption between cloud and fog nodes,the computing efficiency of Internet of Things(IoT)data fusion is studied,and a high-dimensional data compression algorithm for IoT based on fog computing is proposed.Based on the space-time compression mechanism,an offline and real-time hybrid data analysis architecture for fog computing is established,which fully considers the space-time correlation of high-dimensional data and the autocorrelation of the observation matrix.A Kronecker joint observation matrix is constructed,which is solved by the optimization theory to obtain the optimal value.The public data set by the Intel Lab Data are used to verify the effectiveness of the proposed algorithm.Experiment results show that the observation matrix optimization compression mechanism based on the fog calculation has good convergence and lower compression error compared with the EDAM algorithm and the LMDRS algorithm.The relative reconstruction error of the proposed algorithm is the smallest under the same compression ratio.

关 键 词：物联网 雾计算 观测矩阵 数据压缩 高维度数据 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]