物联网传感器数据最近邻-点拓扑目标融合  

Nearest Neighbor Point Topology Target Fusion of Internet of Things Sensor Data

作　　者：李宗达 周曦国[1] LI Zong-da;ZHOU Xi-guo(Computer and Information Engineering College,Tianjin Normal University,Tianjin 300387,China)

机构地区：[1]天津师范大学计算机与信息工程学院,天津300387

出　　处：《计算机仿真》2020年第11期304-307,343,共5页Computer Simulation

基　　金：天津市应用基础研究计划项目重点项目(19JCZDJC35100)。

摘　　要：物联网传感器在执行任务时,由于节点数量巨大,产生大量冗余信息,造成通信带宽浪费等问题。基于此,提出物联网传感器数据最近邻-点拓扑目标融合方法。将置信距离测度当作融合度,通过置信距离矩阵与关系矩阵确定最佳融合数据;利用归一化方法对数据采样矩阵预处理,去除冗余数据;分析传感器数据融合基本原理,结合传感器噪声特点,构建融合模型,定义相对系数与关联系数,计算节点之间最大相似度;通过最近邻-点拓扑目标,选择最佳数据进行初始化处理,根据其它数据记录实现数据融合。仿真结果表明,所提方法数据融合的准确度较高,有效降低了能量消耗,有助于物联网传感器数据的融合。Due to a huge number of nodes in IoT sensor,a lot of redundant information is generated,resulting in communication bandwidth waste.On this basis,a method to fuse nearest neighbor-point topological targets in IoT sensor data was proposed.The confidence distance measures were taken as the fusion degree.The best fusion data was determined by confidence distance matrix and relation matrix.The normalization method was used to process the data sampling matrix and thus to remove redundant data.The basic principle of sensor data fusion was analyzed.Based on the characteristics of sensor noise,the fusion model was constructed.The relative coefficient and the correlation coefficient were defined.The maximum similarity between nodes was calculated.Through the nearest neighbor-point topological target,the best data was initialized,and then the data fusion was realized.Simulation results show that the proposed method has high accuracy in data fusion,which effectively reduces the energy consumption,so it is conducive to the data fusion of sensors in the Internet of things.

关 键 词：物联网传感器 数据融合 最近邻-点拓扑 置信距离 归一化 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]