Actuator fault estimation with disturbance decoupling

The paper deals with actuator fault estimation with disturbance decoupling that makes it possible to diagnose actuator faults. To achieve reliable fault estimation, unknown input decoupling is combined with an observer, leading to a diagnostic observer. As a result, the observation error equation for the diagnostic observer does not depend on the effect of unknown inputs and is used to construct a diagnostic algorithm that provides an estimate of the gains of actuators; this can be obtained directly via the augmented error technique. This approach is compared with the design previously reported in the simulation. The simulation results indicate that the proposed algorithm is more realistic, in the sense that better decoupling properties can be assured without knowledge about unknown inputs, and it is potentially useful in the development of a fault-tolerant control system