基于深度自编码网络的实验室WSN数据压缩算法  

Laboratory WSN data compression algorithm based on deep auto-encoding network

作　　者：时磊 SHI Lei(School of Information and Electronic Engineering,Zhejiang Gongshang University,Hangzhou 310018,China)

机构地区：[1]浙江工商大学信息与电子工程学院,杭州310018

出　　处：《沈阳工业大学学报》2022年第2期193-197,共5页Journal of Shenyang University of Technology

基　　金：国家自然科学基金青年项目(61701441);浙江省高校实验室工作研究项目(ZD201908).

摘　　要：针对无线传感器网络节点冗余和数据重叠的问题,提出一种基于深度自编码网络的实验室WSN数据压缩算法来减少传感器通信和计算开销.使用K-means算法将传感器节点划分为不同的簇,使用深度神经网络对相同簇中节点采集的数据进行压缩编码与分类,从而减小数据规模.使用贝叶斯估计算法对压缩编码后的数据特征进行融合.仿真测试与实验结果表明,所提出的算法能明显减少数据量,相比于Megrez算法与E-CPDA算法的数据融合精度分别提升了4.2%和19%,同时能够增加无线传感器网络的生命周期.Aiming at the problem of node redundancy and data overlap in wireless sensor network,a laboratory WSN data compression algorithm based on deep auto-encoding network was proposed to reduce sensor communication and computational overhead.The K-means algorithm was used to divide the sensor nodes into different clusters,and the deep neural network was used to compress,encode and classify the data collected by the nodes in the same cluster,thus the data size got reduced.The Bayesian estimation algorithm was adopted to fuse the compressed and encoded data features.Simulation tests and experimental results show that the as-proposed algorithm can significantly reduce the data amount.Compared with the Megrez and the E-CPDA algorithms,the data fusion accuracy increases by 4.2%and 19%,respectively,and the life cycle of wireless sensor network gets improved.

关 键 词：自编码网络 无线传感器网络 数据压缩 数据融合 贝叶斯估计 节点 生命周期 计算开销 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]