HW VS SW sensor redundancy: Fault detection and isolation observer based approaches for inertial measurement units

In this paper the use of different observer schemes based on Kalman Filtering for the detection and isolation of aircraft abrupt and incipient sensor faults on Inertial Measurement Units (IMUs) is discussed. The possibility of using a dynamic 6DoF model of the aircraft is explored and compared with the use of a purely kinematic model. Both the possibilities are investigated assuming that two IMUs are available on board, and the analytic redundancy provided by the observers is used to vote the healthy one, when a fault occurs on accelerometers, gyros or magnetometers. The proposed schemes are applied to simulated flight data of a General Aviation aircraft generated in the presence of disturbances and uncertainties. Preliminary experimental results using two low cost IMUs are also shown for possible applications to improve safety and reliability of small Unmanned Air Vehicles.