Dual-Mode deadbeat H2 FIR filtering for discrete-Time systems

Abstract

A new receding horizon (RH) finite impulse response (FIR) filter called the dual-mode deadbeat H2 FIR filter (DMDH2FF) is proposed for state estimation in discrete linear time-invariant (LTI) systems under external disturbances. The DMDH2FF is derived to minimize the H2 norm subject to the deadbeat condition in the disturbance-to-estimate channel, and it can manage the horizon size in two operation modes. The DMDH2FF has better robustness against temporary model uncertainty compared with a conventional H2 infinite impulse response filter. The performance of the DMDH2FF is demonstrated by an F-404 turbofan engine simulation. © 2021 Elsevier B.V.