改进粗糙集理论的无线传感器网络静态节点分类算法  

Static Node Classification Algorithm Based on Improved Rough Set Theory in Wireless Sensor Networks

作　　者：王晓艳 张楠 WANG Xiaoyan;ZHANG Nan(College of Computer Engineering,Shanxi Vocational University of Engineering Science and Technology,Jinzhong Shanxi 030619,China;School of Mechanical and Electrical Engineering,Shanxi Datong University,Datong Shanxi 037003,China)

机构地区：[1]山西工程科技职业大学计算机工程学院,山西晋中030619 [2]山西大同大学机电工程学院,山西大同037003

出　　处：《传感技术学报》2023年第8期1310-1315,共6页Chinese Journal of Sensors and Actuators

基　　金：2022年全国煤炭行业教育研究课题项目(ZMZA20220007)。

摘　　要：因无线传感器网络中静态节点包含一定的不确定信息,决策级融合技术很难对数据进行有效刻画,导致网络静态节点分类效果差。提出一种基于改进粗糙集理论的无线传感器网络静态节点分类算法。采用多次采样法对测试结果加权处理,通过加权移动平均算法对邻居节点进行测试,基于节点密度的混合式算法对数据去噪处理。利用错误分类率中的变精度粗糙集构建改进的粗糙集模型,基于改进粗糙集理论提取无线传感器网络静态节点特征,通过支持向量机构建无线传感器网络静态节点分类模型对节点分类处理。仿真结果表明,所提方法的分类正确率始终高于96.852%、节点召回率在97.321%以上、F1值多于97%、能量消耗在12.52 J~90.20 J之间,证明所提算法具有良好的网络静态节点分类性能。Because the static nodes in wireless sensor networks contain some uncertain information,the decision level fusion technology is difficult to effectively describe the data,resulting in poor classification of static nodes.A static node classification algorithm for wire⁃less sensor networks based on improved rough set theory is proposed.The multiple sampling method is used to weight the test results,the weighted moving average algorithm is used to test the neighbor nodes,and the hybrid algorithm based on node density is used to de⁃noise the data.The improved rough set model is constructed by using the variable precision rough set in the error classification rate.Based on the improved rough set theory,the static node characteristics of wireless sensor network are extracted.The static node classifi⁃cation model of wireless sensor network is built by using support vector machine to classify nodes.The simulation results show that the classification accuracy of the proposed method is always higher than 96.852%,the node recall rate is more than 97.321%,the F1 value is more than 97%,and the energy consumption is between 12.52 J and 90.20 J,which proves that the proposed algorithm has good net⁃work static node classification performance.

关 键 词：无线传感器 网络静态节点分类 改进粗糙集理论 加权移动平均算法 变精度粗糙集 

分 类 号：TP393.04[自动化与计算机技术—计算机应用技术]