Issues of Fault Diagnosis and Fault Tolerant Control for Aerospace Systems

This plenary paper presents an active fault tolerant control system for aerospace applications. The design of the diagnosis module relying on the nonlinear geometric approach leads to the derivation of adaptive filters for fault detection and diagnosis. The fault estimates, which are obtained by exploiting different methods, e.g. neural networks, are used for controller reconfiguration. The design scheme leads to fault estimates decoupled from wind and aerodynamic disturbances in case of aircraft and spacecraft, respectively, thus allowing to achieve very good robustness properties. Moreover, when the wind components are considered as faults, the approach can be applied also for designing an autopilot with wind rejection features. The effectiveness of the designed approach is verified in simulation by considering different examples of aircraft and spacecraft in several conditions and in the presence of faults, turbulence, measurement noise, and modelling errors.