全局最优联邦滤波器信息分配原则

考虑到故障隔离后故障信息的影响一般需要一段时间才能够消除,研究了子系统发生传感器故障条件下的联邦滤波器信息分配问题．基于使故障量对无故障子系统的影响最小化考虑,提出了一种ε信息分配准则,并给出了按给定的故障衰减系数进行调整的信息分配策略．根据本文方法可使其他无故障子滤波器具有较强的抗干扰能力,从而使重构的联邦滤波器快速输出不含故障信息的融合结果．理论分析和仿真说明该方法是可行的．

Information-sharing scheme for federatedfilter with optimality

Because the fault information's influence on the federated filter reconstructed with the HLFs (healthy local filters) can not be removed immediately after the fault isolation, the problem of information-sharing of the federated filter with sensor failures in some local filters is investigated. For improving the ability of reconstruction of the federated filter, an information-sharing criterion under e-restraint condition is theoretically proposed by minimizing the contamination on the HLFs from the faulty local filters, along with as an adaptive adjustment scheme against sensor failures for information division under a given fault attenuation coefficient e. Theoretical analysis and simulation show that, by this method, the HLFs will achieve strong resistance to the contamination before the failures are detected, enabling the reconstructed federated filter to promptly yield a globally optimal solution not containing the fault information.