无线传感器网络自适应预测加权数据融合算法.docx

第30卷第5期2017年5月

传感技术学报

CHINESEJOURNALOFSENSORSANDACTUATORS

Vol．30No．5May2017

AdaptiveForecastWeightingDataFusionAlgorithmforWirelessSensorNetwork?

YUXiuwu1,2?,FANFeisheng1,ZHOULixing1,2,ZHANGFeng1,2

(1.EnvironmentalProtectionandSafetyEngineeringInstitute,UniversityofSouthChina,HengyangHu’nan421001,China;

2.TheStateKeyLaboratoryofSafetyandHealthforMetalMines,MaanshanAnhui243000,China)

Abstract:Inordertoimprovethereliabilityandextendlifetimeofwirelesssensornetwork(WSN)monitoringsystem,anadaptiveforecastweightingdatafusion(AFWDF)algorithmbasedonclusteringisproposed．AFWDFbuildsaforecastmodelbasedonthetemporalcorrelationofdata．Thesourcenodeandtheclusterheadadaptivelyadjustthepredictionmodelparametersbyusingthechangetrendoftheearlymonitoringdatatopredictthelaterda-ta．Thesourcenodeextractstheeigenvaluesandeliminatestheabnormalvaluesbycomparingthepredictedvalueswiththemeasuredvalues．Theclusterheadrestoresthemonitoredvaluesandcalculatesthereliabilityandweightofthemonitoredvaluestofusethethem．Throughtheperformanceanalysisandsimulation,itisconcludedthattheAFWDFalgorithmhashighreliability,andthenetworklifecycleisabout15%higherthanSAEMDAandBPNDAinthesimulationenvironment.

Keywords:wirelesssensornetwork;datafusion;adaptiveforecastmodel;environmentalmonitoring;reliabilityEEACC:7230doi:10.3969/j.issn.1004-1699.2017.05.023

无线传感器网络自适应预测加权数据融合算法?

余修武1,2?,范飞生1,周利兴1,2,张枫1,2

(1．南华大学环境保护与安全工程学院,湖南衡阳421001;2．金属矿山安全与健康国家重点实验室,安徽马鞍山243000)

摘要:为提高无线传感器网络监测系统的可靠性及寿命,提出了一种基于分簇的自适应的预测加权数据融合(AFWDF)算法。AFWDF算法依据数据在时间上的相关性,建立预测模型。源节点与簇头利用前期监测数据的变化态势自适应调整预测模型参数对后期数据进行预测,源节点通过预测值与测量值比较提取特征值和剔除异常值,簇头根据特征值和预测值还原监测值,并计算监测值可信度和权重进行加权数据融合。通过性能分析及仿真,得出AFWDF可靠性较高,且在模拟环境下网络寿命周期比SAEMDA和BPNDA算法提高了15%左右。

关键词:无线传感器网络;数据融合;自适应预测模型;环境监测;可信度

中图分类号:TP393