Reduction of propagated attitude estimation errors using orbital rate coupling in a Kalman filter formulation

Abstract

Earth-pointing satellites with moderate attitude determination accuracy requirements typically use earth sensors and sun sensors for updating the onboard attitude estimate. The earth sensors provide continuous roll and pitch information, however, due to field-ofview limitations and eclipses, the sun sensors can provide yaw updates for only part of each orbit. Propagation of the yaw attitude estimate during the periods of no sun updates can result in a large uncertainty in the estimate. For earth-pointing satellites in a low earth orbit, in which orbital rate is relatively large, this uncertainty can be reduced by modeling in the onboard Kalman filter the roll/yaw coupling inherent in the kinematics of an earth-pointing satellite. An analytical approach to predicting the propagated uncertainty in the yaw estimate is presented herein, and is validated with simulation results. It is shown that by including orbital rate coupling in the dynamics modeling, a significant improvement in yaw estimation accuracy can be achieved during those periods when no yaw updates are available.