摘要：Abstract This paper addresses the problem of detecting and isolating faults in the angle-of-attack (AoA) measurements, when the faulty sensors outnumber the healthy ones. Such a scenario hinders the applicability of heuristic-based techniques, such as voting schemes. Rather than including additional sensors, which would increase payload and system complexity, the solution adopted herein takes advantage of the knowledge on the dynamics of the aircraft, modeled by a linear fractional transformation (LFT), to detect and isolate faults. This approach is used to minimize the impact of the faulty sensors on the control performance of the vehicle, by feeding the controller with the compensated measurement. An integrated design solution is adopted based on μ-synthesis, thus directly addressing the issue of model uncertainty. Simulation results are presented obtained from a very high-fidelity Airbus industrial simulator.
