Committee:  Prof. Kerri Cahoy (Chair), Prof. David Miller, Dr. George Ricker (MIT Kavli Institute), Dr. Roland Vanderspek (MIT Kavli Institute)

Abstract: The Transiting Exoplanet Survey Satellite (TESS) is an MIT-led, NASA-funded Explorer-class planet finder with the primary mission of detecting transiting exoplanets in a 2-year all-sky survey. The TESS instrument consists of four wide-field optical cameras, mounted in a stacked configuration. During science operations, TESS uses the instrument cameras in the loop for attitude determination as part of the fine-pointing system in order to achieve the desired photometric precision of the mission. In this work, we present our approach toward improving and quantifying the fine-pointing performance of TESS and assessing the impact of pointing errors on the overall photometric precision of the mission. First, a guide-star selection method was developed to generate a set of desirable stars for guidance during any arbitrary observation sector based on stellar properties and their proximities to neighboring objects. Next, a comprehensive testing and validation framework was developed to assess the attitude determination flight software as well as to quantify the attitude determination performance of the instrument cameras during key mission scenarios. This framework allows the attitude determination system to be significantly improved, leading to a reduction in open-loop attitude errors by more than 65%. The final attitude determination performance was estimated to meet all relevant open-loop pointing requirements with margin. To assess the closed-loop fine-pointing performance of the system, multiple mission scenarios were simulated and analyzed using a comprehensive framework including both instrument performance as well as spacecraft control and dynamics, showing that the relevant closed-loop pointing requirements at multiple time scales are met with margin. The performance of the system over longer time scales and in temporary camera unavailability periods was also quantified through end-to-end simulations and is in agreement with analytical predictions. Finally, a high-fidelity simulation and analysis framework was developed to assess the photometric precision of the system, including realistic optical responses of the cameras, major photometry noise processes, and expected fine-pointing errors. The simulation results show that with basic co-trending techniques, the impact of pointing errors on science data can be significantly reduced, resulting in shot-noise-limited stellar photometry signals over the magnitude range of interest.