Advanced fault diagnosis techniques in aerospace systems

Discusses the problem of instrument (sensor) fault detection (IFD) in aerospace systems using analytical redundancy. The general framework of model-based fault detection, identification, and accommodation is outlined, and the state of the art in the field is discussed. Emphasis is placed on observer-based residual generation with increased robustness to modeling errors and noise. A time-domain and frequency domain solution to the key problem of designing optimally robust observer schemes on the basis of the unknown input observer theory is described, and the potential of this technique is demonstrated using the example of a flight-control system.<
>