无线传感器网络自适应预测加权数据融合算法

发布时间：2018-04-24 19:53

本文选题：无线传感器网络 + 数据融合　；参考：《传感技术学报》2017年05期

【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。
[Abstract]:In order to improve the reliability and lifetime of wireless sensor network monitoring system, a clustering adaptive predictive weighted data fusion algorithm, AFWDF algorithm, is proposed to build a prediction model based on the temporal correlation of the data. The source node and cluster head self-adaptively adjust the prediction model parameters to predict the later data. The source node extracts the eigenvalue and removes the abnormal value by comparing the predicted value with the measured value. The cluster head restores the monitoring value according to the eigenvalue and the predicted value, and calculates the reliability and weight of the monitoring value for weighted data fusion. Through performance analysis and simulation, it is concluded that AFWDF is more reliable, and the network life cycle is about 15% higher than that of SAEMDA and BPNDA algorithm in the simulated environment.
【作者单位】：南华大学环境保护与安全工程学院;金属矿山安全与健康国家重点实验室;
【基金】：金属矿山安全与健康国家重点实验室开放基金项目(2016JSKSSYS04) 湖南省重点研发项目(2015SK2005) 湖南省教育厅科研重点项目(15A161) 江西省自然科学基金项目(20122BAB201050)
【分类号】：TN929.5;TP202;TP212.9

【相似文献】