一种同质传感器的最优加权融合算法研究  

Research on Optimal Weight Fusion Algorithm of Homogeneous Sensors

作　　者：郑良骏 杨金鑫 王志明[1] ZHENG Liang-jun;YANG Jin-xin;WANG Zhi-ming(School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210000,China)

机构地区：[1]南京理工大学机械工程学院,江苏南京210000

出　　处：《仪表技术与传感器》2022年第4期123-126,共4页Instrument Technique and Sensor

基　　金：国家重点研发计划资助项目(2020YFC1522204,2020YFC1523002)。

摘　　要：为了满足同质传感器采集数据的可靠性需求,提升算法的效率,提出了一种改进的最优加权融合算法。基于3sigma准则,使用估计标准差代替实际标准差剔除粗大误差值,从而提高数据融合的可靠性。利用该算法对一氧化碳的原始数据进行了预处理,结果表明与格拉布斯准则、狄克逊准则等相关算法相比,新算法对粗大误差的剔除更准确,效率更高,减小了待融合数据的标准差。In order to meet the reliability requirement of homogeneous sensor data acquisition and improve the efficiency of the algorithm,the optimal weighted fusion algorithm was proposed.Based on the 3sigma criterion,the estimated standard deviation was used to replace the actual standard deviation to eliminate the outliers to improve the reliability of data fusion.The algorithm was used to preprocess the raw data of carbon monoxide.The results show that compared with the Grubbs criterion and Dixon criterion,the new algorithm is more accurate and efficient in eliminating the outliers and the standard deviation of the fused data can be decreased.

关 键 词：同质传感器 数据融合 最优加权算法 粗大误差 数据预处理 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]