基于奇异值的联邦滤波器信息分配新方法

A New Information-Sharing Method of Federated Filter Based on Singular Value

针对有状态重置结构的联邦滤波器,在分析以往信息分配和重置方法的基础上,提出了一种基于估计协方差阵奇异值动态确定信息分配系数的算法.结合IN S/GPS/DV S/APS组合导航系统的M on te-C arlo数字仿真结果表明:与以往算法和无重置联邦滤波比较,该算法在联邦滤波器的主滤波器信息融合周期长于局部滤波周期时,可有效提高系统状态估计误差精度.

On the basis of analyzing former information-sharing and resetting methods, this paper proposed a new method which determines the information-sharing factors dynamically for federated filter with resetting configuration based on singular value of the covariance matrix of the estimated errors. The Monte-Carlo simulation result of an INS/GPS/DVS/APS integrated navigation system shows that, compared with the former methods and the no-reset federated filter, the proposed method can effectively improve the precision of the estimated errors when the information-fusing period of the chief federated filter is longer than the filtering period of the local filter.