Fault Diagnosis and Fault-Tolerant Control for Avionic Systems

This paper addresses the development of an active fault tolerant control scheme for avionic systems. The methodology is applied to an aircraft longitudinal autopilot taking into account possible faults on the aircraft actuators. The key feature of the proposed control relies on its active characteristics, as the fault diagnosis strategy is based on a robust estimate of the fault signals that are compensated. The design method uses an intelligent data–driven scheme via a fuzzy modelling and identification procedure, which derives these adaptive filters with disturbance decoupling features. The work shows that these fault estimates can be used for fault accommodation. In particular, the fuzzy approach proposed in the paper provides the reconstruction of the fault signals that are decoupled from the wind components, and thus applied to the aircraft system. The proposed solutions provide interesting robustness features that are analysed by using a high–fidelity simulator, which is able to include different operating points and realistic actuator faults, turbulence, measurement errors, and the model–reality mismatch.