分散化H^∞鲁棒滤波理论研究  

Decentralized H-infinity Robust Filtering Theory

作　　者：李金梁[1] 吴训忠[1] 张宗麟[1] 

机构地区：[1]空军工程大学工程学院,西安710038

出　　处：《数据采集与处理》2008年第6期635-640,共6页Journal of Data Acquisition and Processing

基　　金：国防科技重点预研基金(51309060302)资助项目

摘　　要：为改善传统分散化卡尔曼滤波器对系统模型和噪声统计特性的过分依赖性,以及解决分散化H-infinity鲁棒滤波算法的复杂性问题,深入研究了分散化H-infinity滤波的算法结构和鲁棒机理,提出一套H-infinity鲁棒信息滤波公式。通过证明,本文公式与S.H.Jin等人提出的相应公式是等价的,但相比之下更加简单、直观。在此基础上,参考分散化卡尔曼滤波算法的思想,提出一套分散化H-infinity鲁棒滤波算法。与传统算法相比,该算法更加简单、实用。仿真结果表明,新算法有效、可行。To improve the conventional decentralized Kalman filter dependence on the reliable plant models and the statistics of the exogenous signals and solve the complexity of decentralized H-infinity robust filtering, the algorithm structure and the robustness mechanism of the decentralized H-infinity robust filtering are researched. A series of H-infinity information filtering formulas is proposed, and the equivalence relation to the coordinate proposed by S. H. Jin etc. in 2001, is proved. Compared with the latter, the former one is more simple and intuitionistic. Based on the information filtering formulas and the decentralized Kalman filtering algorithm, a series of decentralized H-infinity robust filtering algorithms is put forward. The result proves that the algorithm is more simple and practical than the conventional one. Finally, simulations show that the novel algorithm is effective and feasible.

关 键 词：分散化滤波 KREIN空间 H—infinity滤波 增益失调因子 信息滤波 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]