This paper investigates the attitude estimation capabilities of a debris-removing nanosatellite called deorbiter CubeSat. The spacecraft is designed based on the utilization of commercially-available components with long space heritage, which are embedded in an eight-unit form factor. The attitude estimation machinery employed in this work is a discrete-time, quaternion-based, extended Kalman filter, which utilizes measurements provided by a three-axis rate sensor, five sun sensors, and a three-axis magnetometer. To obtain a linear state-space model, gravity gradient and magnetic disturbance torques are included in the plant model, and the model is linearized with respect to the process noise and the states, namely the inertial angular velocities and the quaternions. measurements noises are modelled based on zero-mean Gaussian distributions, and are quantified based on the performance of the state-of-the-art, commercial-of-the-shelf devices. A Monte Carlo simulation is created to analyze the performance of the estimator against various initial angular velocities and quaternions, both in the sunlit and in the eclipsed portions of the orbit. In light of the results, the accuracy of the deorbiter CubeSat’s attitude knowledge is discussed.
ISBN för värdpublikation: 978-1-5386-9448-0