Globally optimal distributed Kalman filtering fusion  

作　　者：SHEN XiaoJing LUO YingTing ZHU YunMin SONG EnBin 

机构地区：[1]College of Mathematics,Sichuan University,Chengdu 610064,China

出　　处：《Science China(Information Sciences)》2012年第3期512-529,共18页中国科学（信息科学）（英文版）

基　　金：supported in part by National Natural Science Foundation of China (Grant Nos. 60874107,60934009, 60901037, 61004138)

摘　　要：The goal of this paper is to give a survey of the previous works on the globally optimal distributed Kalman filtering fusion with classical and nonclassical dynamic systems. Then, we summarize some of our recent results on nonclassical and unideal dynamic systems, including dynamic systems with feedback and cross-correlated sensor measurement noises, dynamic systems with random parameter matrices, and dynamic systems with out-of-sequence or asynchronous measurements. The global optimality in this paper means that the distributed Kalman filtering fusion is exactly equal to the corresponding centralized optimal Kalman filtering fusion. Therefore, not only all of the proposed fusion algorithms here are distributed, but performance as good as that of the corresponding optimal centralized fusion algorithms is achieved. There also exist many papers for other fusion optimMity （e.g., the optimal convex linear estimation/compression fusion） discussion, which are not involved in this paper.The goal of this paper is to give a survey of the previous works on the globally optimal distributed Kalman filtering fusion with classical and nonclassical dynamic systems. Then, we summarize some of our recent results on nonclassical and unideal dynamic systems, including dynamic systems with feedback and cross-correlated sensor measurement noises, dynamic systems with random parameter matrices, and dynamic systems with out-of-sequence or asynchronous measurements. The global optimality in this paper means that the distributed Kalman filtering fusion is exactly equal to the corresponding centralized optimal Kalman filtering fusion. Therefore, not only all of the proposed fusion algorithms here are distributed, but performance as good as that of the corresponding optimal centralized fusion algorithms is achieved. There also exist many papers for other fusion optimMity （e.g., the optimal convex linear estimation/compression fusion） discussion, which are not involved in this paper.

关 键 词：Kalman filtering distributed estimation fusion FEEDBACK cross-correlated sensor measurementnoises random Kalman filtering out-of-sequence measurements 

分 类 号：O211.64[理学—概率论与数理统计] O224[理学—数学]