基于分离协方差交叉的全局反馈航迹融合  

Track-to-track fusion with global feedback based on split covariance intersection

作　　者：黄珏[1] 颜冰[1] 陈浩文[2] 郭虎生[3] 

机构地区：[1]海军工程大学兵器工程系 [2]国防科技大学电子工程学院 [3]中国人民解放军63981部队

出　　处：《华中科技大学学报（自然科学版）》2016年第1期52-55,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金青年基金资助项目(61302142)

摘　　要：针对多传感器数据融合中互协方差未知的问题,提出了一种基于分离协方差交叉的全局反馈航迹融合方法,将误差协方差矩阵分离为相关信息误差协方差矩阵和独立信息误差协方差矩阵.在此基础上,提出了分离形式的卡尔曼滤波算法,并设计了全局信息反馈的分布式融合结构,克服了简单凸组合融合算法和协方差交叉算法中将所有的信息一起考虑的缺陷.仿真实验表明:本文算法比简单凸组合融合算法和协方差交叉融合方法具有更小的均方误差.For multi-sensor data fusion with unknown cross-covariance,a track-to-track fusion with global feedback based on covariance intersection method was proposed.This method separates the whole covariance into two parts,one is the correlate information error covariance and another is independent information error covariance.Based on this,the form of the split Kalman filter was proposed.Further,a distributed fusion structure with global information feedback was designed.This method overcomes the shortcoming of that the whole information was kneaded together in the method of the simple convex combination fusion or the traditional covariance intersection fusion.Last,a simulation example showed that this method has a smaller mean square error than that two methods.

关 键 词：分离协方差交叉 航迹融合 估计误差 全局反馈 卡尔曼滤波 

分 类 号：TN911.7[电子电信—通信与信息系统] O211.64[电子电信—信息与通信工程]