State estimation of cooperative satellites for on-orbit assembly and servicing of spacecraft

Thesis: S.M., Massachusetts Institute of Technology, Department of Aeronautics and Astronautics, 2017.
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 153-157).
The development of robust and routine execution of autonomous space-based proximity operations is a critical need for the future of space exploration and space-based business enterprise. One application of this host of activities, which includes rendezvous, capture, and docking, is on-orbit assembly and servicing of spacecraft. It is believed that the maturation of this technology could usher in a new era of space technology featuring modular construction of large spacecraft or habitats for exploration and tourism, assembly of large-aperture space telescopes unconstrained by launch vehicle size, and reconfigurable structures for mission adaptability. Furthermore, this technology could extend to capture and repair high asset spacecraft by replacing modular components, all without the need to risk human lives. This thesis seeks to contribute to the development of this technology by focusing on one of its most critical aspects: robust state estimation between the autonomous agents. Several estimation frameworks exist that can be applied. However, two state estimators were specifically chosen, implemented, verified, and validated for reasons discussed in the text. First, a practical method of implementation of an Unscented Kalman Filter for two active, cooperative, autonomously docking satellites that overcomes latency issues from low frequency vision-based relative-pose measurements is presented. Second, a factor graph based incremental smoothing estimator for the same application is implemented, which can be shown to provide robustness to several failures characteristic of the filtering framework. A detailed analysis enumerating the strengths and weaknesses of the two frameworks is provided, as well as the verification and validation of the two estimators via the SPHERES testbed from both a 3-DOF planar air bearing facility and the playback of data sets collected from the International Space Station 6-DOF test environment.
by William D. Sanchez.
S.M.