Fault Detection Using Differential Flatness in Flight Guidance Systems

In this paper, flight guidance dynamics are shown to be implicit differentially flat with respect to the inertial position of an aircraft. This proves the existence of a set of relations between these flat outputs and the state variables representative of flight guidance dynamics and between these flat outputs and the basic inputs to flight guidance dynamics. A neural network is introduced to obtain, from the actual trajectory, nominal flight parameters which can be compared with actual values to detect abnormal behaviour