Kalman filter stationarity analysis in the problem of INS errors estimation

Abstract

Kalman filter has long been used in aircraft integrated navigation systems as the main instrument of estimating navigation errors. Its main advantage over simpler state estimators is the ability to adjust filter's gain according to observable signal instability. But it is possible to use simple observers if filters gain is stationary regardless of aircraft maneuvers. In this paper the model of navigation errors dynamics is considered. This model is used in Kalman filer in integrated system simulation. Frequency domain representation of vehicle dynamics is used in this simulation to investigate filter stationarity. The results show that filter gain is stationary regardless of motion dynamics.