Author: Hugo Miguel Martins Marques
Instituto Superior Técnico, Lisboa (Portugal)

The Attitude Determination and Control System (ADCS) of a nanosatellite is a key subsystem to provide precise attitude knowledge and pointing for the on-board payload and necessary maneuvers. Its design has serious constraints in terms of mass, volume, size, cost and power. The main goal of this dissertation is to provide the NANOSTAR project with a grounded study in terms of attitude determination algorithms and sensors that can be employed in the missions designed under the scope of the project. For that, a simulation platform that realistically describes the nanosatellite environment, allowing orbit generation and propagation, as well as data creation to feed the ADCS was developed.

Then, three representative attitude determination algorithms, namely the Quaternion Estimator (QUEST), the Multiplicative Extended Kalman Filter (MEKF), and a recently developed Globally Exponentially Stable Cascade Attitude Nonlinear Observer, were studied and implemented on the platform developed, using vector measurements provided by a star tracker, Sun sensor, magnetometer and rate gyroscope. Finally, the comparison of the three algorithms in terms of computational resources efficiency, steady-state performance and performance in the case of faults is done, using realistic simulation scenarios. The results obtained provide meaningful insight on the advantages, disadvantages, complexity, computational resources efficiency and performance of the three algorithms,
providing the project with a grounded analysis that can be used for future decision making in terms of
the ADCS design.