LPV sliding mode observers for sensor fault reconstruction with erroneous scheduling parameter measurements

In this paper, a linear parameter-varying sliding mode observer scheme is proposed to reconstruct sensor faults and simultaneously estimate the real value of an erroneous scheduling parameter in the baseline controller, so as to achieve active fault tolerant control. In this scheme, an adaptive algorithm is developed, driven by the so-called `equivalent output error injection´ signal associated with the reduced-order sliding motion. A sufficient condition is given which ensures asymptotic stability of the state estimation error system, wherein both the state estimation errors and the estimation error between the nominal scheduling parameter and its estimate converge to zero. The scheme is then applied to a RECONFIGURE benchmark to deal with a scenario involving the simultaneous total loss of Calibrated Airspeed and Angle of Attack measurements.