Zonotopic Fault Diagnosis Method for Linear Systems with Unknown Input Set-membership Filters

Abstract

This work develops a novel zonotope-based fault diagnosis algorithm for uncertain systems. First, an unknown input zonotopic filter for each actuator is set up to acquire the measurement output set, by regarding the actuator fault as unknown input and estimating the output boundary by zonotopic analysis. Second, only its filter is decoupled when an actuator has a fault, while other filters are affected by the fault. Then, the strategy for fault diagnosis based on a bank of zonotopic filters is put forward. Finally, to verify the feasibility of the suggested approach, numerical simulations of a flight longitudinal motion control system are demonstrated.