Seminars

Advancements in the fields of spacecraft navigation and autonomy are needed to ensure future vehicles can complete their missions safely and efficiently. This dissertation proposes several advancements in this domain. First, a process is outlined for mapping an unknown environment using only
monocular camera images. It is proposed that the Modified-Half-Normal distribution can represent map uncertainty more faithfully than the usual Gaussian. The usefulness of this uncertainty model is demonstrated with synthetic images from a spacecraft proximity operations scenario. Second, a variable- resolution system for eƯiciently mapping terrain and selecting landing sites is developed. This system is proven with flight-test data from a rocket-powered vehicle. Finally, a new Kalman filter is derived for spacecraft with redundant three-axis gyroscopes. The filter fuses all gyroscope measurements and provides a natural method for detecting and isolating faulty gyroscope axes.