Design of a robust nonlinear attitude estimation algorithm for Space Rider mission

Design of a robust nonlinear attitude estimation algorithm for Space Rider mission An important part of almost every space mission post-launch analysis is the at- titude sensors performance assessment. Launch vibrations, orbit insertion and a harsh operating environment could degrade the sensors data reliability. Even mounting errors during vehicle assembly and integration could occur, mining the mission success. Generally, spacecraft attitude sensors require extensive in-flight calibration during their operational life to ensure pointing requirements are satis- fied. Continuous on-board calibration also provides a means for fault prediction and detection trough parameters tracking. Star trackers are among the most accurate instruments to estimate a spacecraft orientation in space, achieving accuracies to the arc-second range in the boresight pointing direction. Effectiveness and reliability have made this sensor an irreplace- able component for the attitude determination system of large satellites and, as the technology improves and allows the miniaturization of the equipments, even for the small and micro ones. However, in order to maintain high pointing accuracy during the entire operative life, it’s necessary to account for misalignments, lens distortion and sensor alterations, due by the environmental changes throughout the entire mission envelope. The main objective of this thesis is to exploit the usefulness of spacecraft dynamic modeling for nonlinear attitude state estimation techniques, investigating the feasi- bility of estimation algorithms to assess the star trackers misalignments w.r.t their mounting directions. Although the nonlinear nature of the spacecraft dynamics doesn’t allow for optimal solution, sub-optimal nonlinear state estimation filters are provided. In order to estimate the vehicle states and compensate for the de- grading performances of the inertial sensors, three filters have been implemented: a Kalman filter in its linear and extended formulation and a variable structure ob- server based in the sliding mode. A continuously operating EKF-based calibration filter estimates attitude rate and quaternion orientation, producing optimal attitude solutions, in therms of mini- mum variance, regardless of the attitude motions. Instead Sliding Mode Observers are typically used for the design of attitude and angular velocity determination al- gorithms (routines) to reduce the computational load of traditional nonlinear filters but preserving their accuracy and stability. Space Rider is a reusable unmanned space transportation system, integrated with Vega-C, designed and developed by ESA and partners to provide a regular access to LEO for several space applications. Space Rider inertial parameters and sensors are implemented in a comprehensive framework in Matlab and Simulink environment and Montecarlo simulations have been performed to test the filter’s performances with different initial conditions and scenarios.