THE ANALYSIS OF GPS ATTITUDE DATA FROM A SLOWLY ROTATING, SYMMETRICALGRAVITY GRADIENT SATELLITE

The purpose of this paper is to compare theoretical, attitude dynamica l models for a gravity gradient satellite to empirical results based o n data collected on-board the Air Force RADCAL satellite using a Globa l Positioning System (GPS) receiver. The on-board data is processed us ing a Kalman filter which estimates attitude angles, rates, and distur bance torques. The filter is tuned by examination of the autocorrelati on of the angle residuals. The optimal tuning is obtained when the onc e and twice per orbit periods are minimized in the residuals and when periods shorter than 20 minutes are minimized in the disturbance torqu es. The correlations in the residuals with periods near ten minutes ar e attributed to the integrated GPS multipath from the spacecraft struc ture. The attitude control system works to the design level with the l ibrations controlled to less than 20 degrees in roll and pitch. The fl uctuations in the librational energy are at the level predicted from p re-flight analysis of a hysteresis damper.