Spacecraft Integrated Pose Determination with Unknown Noise Covariances

A novel robust unscented Kalman filter (UKF) is developed for orbiting spacecraft pose determination without knowing prior noise statistical characteristics by using dual quaternion formulation. First, an adjustment coefficient multiplied by a predesigned positive definite matrix is updated iteratively for the process noise covariance adaptation and the moving window of the measurement innovation are adopted for the measurement noise covariance estimation. Then, the adaptation method is incorporated into the dual-quaternion-based UKF procedure where the dual modified Rodrigues parameters (MRPs) are adopted for sigma points description to avoid violating the quaternion normalization constraint. Finally, the effectiveness and practical usefulness of the developed algorithm are verified via numerical simulation examples.