Sensor fault estimation using LPV sliding mode observers with erroneous scheduling parameters

This paper proposes a linear parameter-varying sliding mode observer for the purpose of simultaneously estimating the system states and reconstructing sensor faults. Furthermore, some of the measured scheduling parameters are also assumed to be unreliable, and the corresponding values used in the observer are adapted to maintain the performance level of the observer. The adaptive algorithm is driven by the 'equivalent output error injection' signal associated with the reduced-order sliding motion. Sufficient conditions are given to ensure asymptotic stability of the state estimation error system, ensuring both the state estimation errors and the estimation errors associated with the scheduling parameters converge to zero. The efficacy of the scheme has been evaluated based upon an industrial high-fidelity aircraft benchmark scenario involving a simultaneous total loss of airspeed and angle of attack measurements. (C) 2018 The Authors. Published by Elsevier Ltd.