Aircraft inertial measurement unit fault identification with application to real flight data

Abstract

An iterated optimal two-stage extended Kalman filter (IOTSEKF), which improves the performance of the optimal two-stage extended Kalman filter (OTSEKF) when dealing with nonlinear systems, is applied to estimate both the system states and the IMU faults. The simulation data are taken from the Aero-Data Model in a Research Environment (ADMIRE) benchmark model. First, the KM2 is proposed for the fault identification of the inertial measurement unit sensor faults. It takes the wind into account, which makes it less sensitive to turbulence compared to the kinetic model (KM1). Second, the fault identification (FI) using the KM2 is tackled by proposing a novel iterated optimal two-stage extended Kalman filter, which improves the performance of the optimal two-stage extended Kalman filter when dealing with nonlinear systems. Finally, the inertial measurement unit (IMU) sensor FI performance of the IOTSEKF using the KM1 and the KM2 is further validated by using the recorded real flight-test data. Results demonstrate the effectiveness of the approaches. The proposed approach using the KM1 and the KM2 can be further extended to fault tolerant control systems, which can provide more accurate information to enhance the safety of the aircraft when there are malfunctions in the IMU sensors.