Robust Strong Tracking Cubature Kalman Filter for Attitude Estimation of Failed Spacecraft

Abstract

The malfunction of attitude control system will lead to continuous attitude tumble of on-orbit spacecraft, which brings two problems to attitude estimation, one is the large initial error, and the other is the status mutation. Focusing on the above two issues, a robust strong tracking cubature Kalman filter (RSTCKF) is proposed in this paper to obtain the high-precision and real-time attitude knowledge of failed spacecraft. On the one hand, the multiplicative quaternion is adopted in the whole filter process instead of the additive one to keep the quaternion normalized, which increases the numerical stability of CKF; on the other hand, the time-varying multiple fading factors are introduced to adjust the different channels of prediction error covariance matrix, which strengthens the tracking ability of CKF to status mutation. Numerical simulations verify the effectiveness of the developed algorithm.