Process and Measurement Noise Covariance Tuning in Kalman-Based Estimator Aided by SVD

Process and measurement noise covariance matrices are tuned for an adaptive attitude estimation of a nanosatellite at low Earth orbit based on extended Kalman filter (EKF) that is added by singular value decomposition (SVD) method. The tuning procedure compensates the measurement and process noise covariance variations. The tuning of the R matrix is simply processed in SVD, one of the single-frame methods. The tuning of the Q matrix is defined in the second stage of the Kalman-based estimator design. The tuning rules are run at the same time, so the filter is capable of being robust against initialization errors, system noise uncertainties, and measurement malfunctions without an additional filter design usage.