A direct/functional redundancy scheme for fault detection and isolation on an aircraft

In this paper the problem of detecting and isolating sensor faults on aircraft, in the presence of external disturbances, is considered. A novel nonlinear observer based approach is proposed; this approach exploits the so-called Unknown Input Observer (UIO) theory and guarantees disturbance decoupling with an H∞ performance level. The proposed nonlinear observer has been tested and compared with the classical linear UIO scheme via numerical simulations performed on the model of the small commercial aircraft under consideration; this simulation activity has shown the benefits of the novel approach. The proposed methodology is finally exploited to design an algorithm which, on the basis of the hardware and software measurements, decides if and where a fault has occurred, isolates the faulted signal, and outputs the corresponding healthy signal. Tests performed on the flight simulator of the National Aerospace Laboratory in Amsterdam show that this algorithm improves the performance of a classical duplex hardware system.