A Covariance Matching-Based Adaptive EKF for Nanosatellite Attitude Estimation

This study discusses and compares the performance of R-adaptive extended Kalman filters (EKF) with different covariance matching techniques for a nanosatellite attitude estimation. A non-traditional approach is used for the estimation process where the TRIAD and an EKF are integrated to reduce the computational load. In order to make the EKF adaptive, covariance matching techniques are used with single scaling factor (SSF), multiple scaling factors (MSFs), and fading factors (FFs), which is an alternative approach to MSFs. To compare the performance of the proposed algorithms, one simulation is performed where a noise increment is applied to the x-axis magnetometer. As a result of the simulation, it is seen that MSF and FF approaches are superior to the SSF approach. On the other hand, no differences in performance are observed between the MSF and FF approaches.