迭代异方差估计及其在多传感器数据融合中的应用  

Iterative Heteroscedastic Variance Estimation with Its Applications for Multisensor Data Fusion

作　　者：赵晋[1] 张建秋[1] 高羽[1] 

机构地区：[1]复旦大学电子工程系,上海200433

出　　处：《电子学报》2008年第10期1938-1943,共6页Acta Electronica Sinica

基　　金：国家自然科学基金(No.60572023)

摘　　要：本文提出一种适应任意噪声分布的迭代渐近无偏估计异方差的方法,它能对多个不同测量噪声的方差进行估计.不同于传统的异方差估计算法,本文提出的迭代异方差估计,可在不损失估计精度和减少运算量的前提下,对多个不同的测量噪声方差进行捕获和跟踪.在多传感器数据融合中的应用结果表明:本文提出的方法具有估计稳定性好、运算简单和具有较强的鲁棒性等优点,仿真和实验的结果均证明了提出方法的有效性和可行性.In this paper, an iterative asymptotic unbiased heteroscedastic variance estimation method, suitable for the measurement noise with arbitrary possibility distributions, is presented. The method can estimate the variances of several different measurement noises at the same time. Unlike traditional heteroscedastic variances estimation algorithms, the proposed iterative estimation method can capture and track the vary variances of the measurement noises with less computation while its estimation accuracy is kept. Its application results on multisensor data fusion show that the presented algorithm is advantageous to the estimation stability, computation simplicity, and good robustness. Both the simulation and experiment results verify the effectiveness of the proposed method.

关 键 词：迭代异方差估计 最小均方误差 多传感器数据融合 加权最小二乘 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]