Memory fading unscented Kalman filter and its application in navigation by stellar refraction

This paper emphasizes improving the navigation performance, when refraction starlight is used to compute the position and velocity of a satellite in unscented Kalman filter. In order to avoid the impact of inaccurate measurement information, a memory fading unscented Kalman filter is proposed. Influenced by the fading factor, the refreshment of navigation results in unscented Kalman filter becomes a dynamic process, in which the new measurement information plays a more important role. Besides, the memory length of measurement information is limited in a time interval to completely isolate the old measurement information. Theoretical analyses and simulation results show that the memory fading unscented Kalman filter is obviously better than the standard unscented Kalman filter in reducing navigation errors.