无线传感器网络中数据密度相关度融合算法  

Data density correlation degree based aggregation algorithm in wireless sensor network

作　　者：袁飞[1,2] 詹宜巨[3] 王永华[4] 

机构地区：[1]广东技术师范学院自动化学院,广东广州510665 [2]中山大学信息科学与技术学院,广东广州510006 [3]中山大学工学院,广东广州510006 [4]广东工业大学自动化学院,广东广州510006

出　　处：《控制理论与应用》2014年第11期1568-1573,共6页Control Theory & Applications

基　　金：国家自然科学基金资助项目(61071038);广东省教育部产学研结合重点资助项目(2011A090200128)

摘　　要：无线传感器网络节点部署在复杂环境时,节点间相关性无法通过节点间距离来准确描述.为了克服该缺陷,本文提出了数据密度相关度公式.该公式反映了节点数据的ε邻域内数据的聚集程度,也反映了该节点数据相对其ε邻域内数据的相对位置.同时,将数据密度相关度公式应用到代表式数据融合算法中,提出了数据密度相关度融合算法.该融合算法得到的相关区域具有相关区域内节点数据相关度大,相关区域问节点数据相关度小的优点.仿真实验结果表明了该融合算法在数据准确性和能耗方面较基于α-局部空间数据融合算法和基于皮尔森相关系数的数据融合算法优越.Distance between sensor nodes can＇t reflect their correlation degree in wireless sensor network（WSN） as the sensor nodes are deployed in complex environment.In order to resolve this drawback,data density correlation degree（DDCD） is proposed in this paper.The DDCD is a spatial correlation measurement of a sensor node＇s data to its neighbor nodes＇ data.It could reflect the concentration degree of neighbor nodes＇ data.As well,it could describe relative position of a sensor node＇s data to its e-neighborhood data.Based on this correlation degree,DDCD aggregation algorithm is presented to highlight that the representative data has a low distortion on the represented data in WSN.Additionally,simulation experiments with a real dataset are presented to evaluate the performance of the DDCD aggregation algorithm.The experimental results show that the resulting representative data achieved by DDCD aggregation algorithm have a lower data distortion than those achieved by the Pearson correlation coefficient based data aggregation algorithm or α-local spatial data aggregation algorithm.Moreover,the energy consuming of DDCD aggregation algorithm is less than those of the other two data aggregation algorithms.

关 键 词：数据融合 无线传感器网络 相关性 数据密度 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置] TN929.5[自动化与计算机技术—控制科学与工程]