一种传感器网络时空融合模型  

Spatial-temporal Data Fusion for Sensor Networks

作　　者：康健[1] 左宪章[2] 唐力伟[1] 张西红[3] 李浩 

机构地区：[1]军械工程学院火炮工程系,河北石家庄050003 [2]军械工程学院电气工程系,河北石家庄050003 [3]军械工程学院计算机工程系,河北石家庄050003 [4]63880部队,河南洛阳471003

出　　处：《军械工程学院学报》2010年第3期63-66,共4页Journal of Ordnance Engineering College

基　　金：国家自然科学基金资助项目（60672143）

摘　　要：针对多传感器间存在时空相关性和先验知识未知的情况,提出了一种基于核偏最小二乘最优加权的传感器网络时空融合模型。首先用核偏最小二乘法对每个传感器在不同时刻的测量值进行融合,然后将各个传感器同时刻的估计值进行空间最优融合,从而得到被测参量的最终融合估计。该模型还能排除那些取值远离样本点集合平均水平的奇异点。理论分析和融合实验表明,该模型能取得较高的融合精度,并能显著节省网络能量。To solve the problem that spatial-temporal correlation between the multi-sensors and the prior knowledge is unknown, a new spatial-temporal data fusion model for sensor networks is presented. In this model,the data that is monitored by a sensor at different times are fused first in time order based on kernel partial least squares,and then the temporal fusion data are fused in space based on optimal weighted fusion. The model also can eliminate the overlapping and meaningless information. Theoretical analysis and experimental results show that the model can achieve high precision fusion and flexibility. It also can observably reduce the number of transmissions in sensor networks to save energy.

关 键 词：核函数 传感器网络 数据融合 偏最小二乘 方差估计 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]