A novel robust unscented Kalman filter for spacecraft attitude estimation

Abstract

The problems of packet dropping and non-Gaussian noise in spacecraft attitude estimation systems due to the influence of complex external environments are challenging topics. To solve this problem, a novel robust unscented Kalman filter (NRUKF) is proposed in this paper. First, an augmented model associated with the packet-dropping variable is constructed by designing the coefficient matrix. Then, a robust filter with a recursive process is designed based on the minimum error entropy criterion with fiducial points (MEE-FP). The prior estimates are given using the propagation equation, and the posterior estimates are given using the fixed-point iteration algorithm. Finally, the experimental results show that the proposed NRUKF has excellent performance when the spacecraft attitude estimation system suffers from packet dropping and non-Gaussian noise.