A Fault-Tolerant Filter for Systems Subject to Stochastic Faults

[摘要] This paper is concerned with state estimation of discrete-time linear systems subject to additive faults represented as inputs to both the state and measurement equations. Particularly, the sequence of fault inputs is assumed to be parameterizable by three fault parameters: fault magnitude, fault instant, and fault mode index. Moreover, these parameters are treated as unknown realizations of random variables defined so as to account for prior knowledge about possible faults. For tackling this problem, a two-stage filter structure is invoked.

[学科分类] 自动化工程

[关键词] Fault-tolerant state estimation ;Filtering theory ;Statistical signal processing ;Bayesian detection