基于分时融合反馈的联邦滤波算法研究  

The Research on the Federated Filter Based on the Time-sharing Feedback and Fusion Principle

作　　者：臧荣春[1] 李红宝[1] 王海龙[2] 

机构地区：[1]航天东方红卫星有限公司,北京100094 [2]中国空间技术研究院载人航天总体部,北京100094

出　　处：《航天控制》2011年第4期14-17,22,共5页Aerospace Control

摘　　要：传统联邦卡尔曼滤波算法的高容错性是以牺牲精度为代价的,为此提出了改进的联邦滤波算法,采用分时融合反馈的联邦滤波信息分配方式,在提高系统容错性的同时兼顾导航精度和实时性,而且无融合反馈阶段延长了故障反应时间,提高了系统对渐变故障的检测能力,取得较为理想的结果。Aiming to solve the conflict between the tolerance and the accuracy, an improved federated filter is proposed and a time-sharing feedback federated filter is presented which is based on the information-sha- ring principle. The algorithm greatly improves the fault tolerance and keeps the accuracy and real-time as well. Thus, it is possible that the gradually changed fault can be accumulated big enough regarding the de- sign of non-feedback phase. Moreover the capability of detecting soft-fault is improved, and the efficiency is proven by the simulation results.

关 键 词：组合导航 联邦滤波 故障隔离与系统重构 

分 类 号：V249.328[航空宇航科学与技术—飞行器设计]