Show simple item record

dc.contributor.authorZhu, Frances
dc.identifier.otherbibid: 11050421
dc.description.abstractTechnology enables space exploration and scientific discovery. At this amazing intersection of time, new software and hardware capabilities give rise to daring robotic exploration and autonomy. Close-proximity operations for spacecraft is a particularly critical portion of any robotic mission that enables many types of maneuvers, such as docking and capture, formation flying, and on-orbit assembly. These dynamic maneuvers then enable different missions, like sample return, spacecraft construction larger than a single rocket faring, and deep-space operations. Commonly, spacecraft dynamic control uses thrusters for position and attitude control, which rely on active sensing and consumable propellant. The development of other dynamic control techniques opens new capabilities and system advantages, and further offers a more diverse technological trade space for system optimization. This research comprehensively investigates the utilization of flux-pinning physics to manipulate spacecraft dynamics. Flux-pinned interfaces differ from conventional dynamic control through its passive and compliant behavior. These unique characteristics are extremely attractive for certain applications, but flux-pinned technology must mature considerably before adoption for spaceflight missions. A dynamic capture and docking maneuver in an upcoming mission concept, Mars Sample Return, motivates the technology design. This body of work as much as possible follows a progression from cradle to grave. A flux-pinning theoretical dynamics model and a system architecture are presented to specify general capabilities of such a spacecraft system. Different analyses on stability, state sensitivity, backwards reachability result from a physics-based dynamics model. An extensive literature review and basic science experiments inform a theoretical dynamics model about the incorporation of physical parameters when simulating realistic dynamics. A series of testbeds enable experimentation and precise investigation of flux-pinned interface capabilities in the context of docking and capture. The testbeds ranged from the simplest expression of dynamics, in a single degree of freedom, to a flight traceable expression, in all six degrees of freedom. Experiments from these testbeds define and characterize system level capabilities specific to flux-pinned capture. Data collected from these experiments then supports development of a predictive dynamics model of the hardware system. Various system identification methods aid in creating a dynamics model that accurately predicts the dynamics observed during experiments. Several objective metrics are considered to evaluate the model fidelity. The types of system identification methods are separated into analytical methods and numerical methods. The analytical method involves parameter estimation in a physics-based model. Numerical methods involve Taylor expansion, bag of functions, symbolic regression, and neural networks. Theoretical extensions towards verification further develops neural network approximation methods, driving at safe, real-time system identification.
dc.rightsAttribution 4.0 International
dc.subjectSpacecraft Dynamics
dc.subjectSpace Systems
dc.subjectMechanical engineering
dc.subjectAerospace engineering
dc.subjectComputer science
dc.subjectsystem identification
dc.typedissertation or thesis Engineering University of Philosophy, Aerospace Engineering
dc.contributor.chairPeck, Mason
dc.contributor.committeeMemberKress Gazit, Hadas
dc.contributor.committeeMemberFerrari, Silvia

Files in this item


This item appears in the following Collection(s)

Show simple item record

Except where otherwise noted, this item's license is described as Attribution 4.0 International