Adaptive Sliding Mode Fault Diagnosis for LPV Systems with Uncertain Scheduling Variables

This paper presents an adaptive sliding mode observer for actuator fault diagnosis of linear parameter-varying (LPV) systems with imperfectly measured scheduling variables due to noisy or faulty measurements. The developed adaptive observer is able to cope with bounded uncertainties and faults without a priori knowledge of their bounds. Moreover, the actuator fault can be estimated from the principle of equivalent control after minimizing the effect of the scheduling variable uncertainties. The performance of the proposed design approach is validated using simulation studies.