基于最小二乘原理多传感器加权数据融合  

A weighted multi-sensor data fusion based on least squares principle

作　　者：赵华哲[1,2] 李强[1,2] 杨家建[1,2] 

机构地区：[1]南昌航空大学信息工程学院,江西南昌330063 [2]南昌航空大学物联网研究所,江西南昌330063

出　　处：《微型机与应用》2013年第12期7-10,共4页Microcomputer & Its Applications

摘　　要：针对目前多传感器数据融合过程中传感器对某一状态量测量时精度较低的问题,提出了基于最小二乘原理的多传感器加权数据融合算法。该方法利用最小二乘原理和方差的遗忘信息,通过均方误差比较,计算出各个传感器的权重之后进行加权融合。该算法既考虑了历时信息的作用,又考虑了环境噪声和新采样值的影响,增强了对环境监测的敏感性。相比同类融合方法,该方法具有较高的精度,最后仿真结果也直观地说明了该方法的有效性。For the difficulty of low precision of multi-sensor fusion result ot measurement data,the paper proposes a welgnteu multi-sensor data fusion algorithm based on the least squares principle.The method uses the principle of least squares and the vari- ance of forgetting information, compares mean square error, calculates each sensor weight and uses a weighted fusion algorithm.The algorithm takes the effect of history data and ambient noise and new sample into account ,enhances sensitivity to noise measure.Fi- nally, the simulation shows that the fusion results have higher precision compared with other methods.

关 键 词：数据融合 多传感器 均方误差 最小二乘 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]