Fault Diagnosis for Nonlinear Uncertain Systems Using Robust/Sliding Mode Observers

The authors investigate the fault diagnosis problem for a class of nonlinear multiple-input multiple-output (MIMO) systems with uncertainty. Under some geometric conditions, the system is transformed into two different subsystems with uncertainty. The first one is in the generalized observer canonical form, which is not affected by actuator/sensor faults. Nonlinear robust observer or sliding mode observer can be constructed for this subsystem depending on different assumptions on the uncertainties. The second subsystem, whose states can be measured, is affected by the faults. The observation scheme is then used for actuator/sensor fault estimation with good accuracy. Finally, a flexible joint robotics example is given to illustrate the efficiency of the proposed methods.