Robust observer based fault diagnosis for a class of nonlinear systems with uncertainty

The fault diagnosis problem for a class of nonlinear multiple-input multiple-output (MIMO) systems with uncertainty is investigated. Under some geometric conditions, the system is transformed into two different subsystems. One is in the generalized observer canonical form which is not affected by actuator/component faults, so the nonlinear robust observer for this subsystem can be constructed. The other whose states can be measured is affected by the faults. The observation scheme is then used for actuator/component fault estimation with certain accuracy. Extension to sensor fault diagnosis is also discussed. Finally, a numerical example is used to illustrate the efficiency of the proposed method.