AESMF based sensor fault diagnosis for RUAVs

An Adaptive Extended Set-Member Filter (AESMF) with the adaptive selection scheme of the filter parameters is incorporated with the nonlinear attitude state estimation equation to build a sensor fault diagnosis system which can provide guaranteed sensor fault detection. Compared with other sensor fault diagnosis systems based on Kalman Filter (KF) or other probability based methods which can just provide a fault probability distribution but not tell the exact result, in this paper, with the advantage of ellipsoid bound of set-member, we try to implement AESMF to tackle this problem and provide the exact fault diagnosis result. The AESMF is incorporated into the navigation system equation and the sensor fault diagnosis method is introduced. Simulations are conducted and the algorithm is compared with the EKF based navigation system, the result demonstrates the improvement of this method.