CI算法在两传感器融合稳态Kalman滤波器中的应用  

Two-sensor Information Fusion Steady-state Kalman Filter Based on CI

作　　者：黄铫[1] 张天骐[1] 刘燕丽[1] 夏淑芳[1] 

机构地区：[1]重庆邮电大学信号与信息处理重庆市重点实验室,重庆400065

出　　处：《弹箭与制导学报》2010年第3期165-168,共4页Journal of Projectiles,Rockets,Missiles and Guidance

基　　金：国家自然科学基金-中物院NSAF联合基金(10776040);国家自然科学基金(60602057);信号与信息处理重庆市重点实验室建设项目(CSTC;2009CA2003);重庆市科委自然科学基金(CSTC;2006BB2373;CSTC;2009BB2287);重庆市教委自然科学基金(KJ060509;KJ080517);重庆邮电大学自然科学基金(A2006-04;A2006-86)资助

摘　　要：针对两传感器稳态Kalman滤波器的信息融合,目前有三种常用的加权分布式融合算法:按标量加权、按对角阵加权和按矩阵加权,但它们都需要得到局部稳态滤波误差互协方差阵后才能计算出融合结果。而协方差交集算法在相关度未知的情况下,也能得到一个改进的估值,因此文中将协方差交集算法应用到两传感器稳态Kalman滤波器的信息融合中,在互协方差阵未知的情况下,此方法也能得到较好的信息融合结果,并通过仿真进行了验证。For information fusion of the two-sensor steady-state Kalman filter, there are three common weighted distributed fusion algorithms： weighted by scalar, weighted by diagonal matrices and weighted by matrices. They all need calculating the local stead- y-state filtering error cross-covariance matrix to obtain results. The covariance intersection algorithm can get an improved valuation in the face of unknown relevance. In this article, the covariance intersection algorithm was applied to the two-sensor information fusion steady-state Kalman filter, in the case of unknown cross-covariance matrix; good information fusion results can be obtained. Finally, an analysis of the experiment results is made and validated.

关 键 词：两传感器 稳态Kalman滤波 协方差交集算法 

分 类 号：TN713[电子电信—电路与系统]