Resilient decentralized filtering of interconnected discrete-time systems

The objective of this paper is to propose new tools for robust and resilient filter design of a class of interconnected discrete-time systems with uncertain function of nonlinear perturbations by exploiting the decentralized information structure constraint. We seek to establish complete LMI-based procedures for linear filtering by basing all the computations at the subsystem level. The robustness property of filter design is assessed in terms of ℓ ∞ performance while the resilience property is evaluated in face of additive gain variations that reflect the imprecision in filter implementation. By a suitable convex analysis, it is shown that both design problems can be formulated as convex minimization problem over linear matrix inequalities (LMIs). Numerical examples are provided to illustrate the effectiveness of the developed design methods.