Accurate Three-Axis Control of Spacecraft with Non-Uniform Mass Distribution
Authors
Annesen, Patrick Munoz ; Mendes Goncalves, Manuela
Term
4. term
Education
Publication year
2023
Submitted on
2023-06-02
Pages
64
Abstract
This thesis, conducted with the Danish satellite manufacturer Space Inventor, investigates attitude control when a satellite has an uneven mass distribution. The company proposed a coordinate transformation to control along the principal axes—the satellite’s natural axes of rotation—to counteract mass imbalance. We implemented this idea using two standard controllers, a PD controller (proportional–derivative) and an LQR (Linear Quadratic Regulator), and compared it with controlling directly in the body frame (coordinates fixed to the spacecraft). The comparison used Monte Carlo simulations (many randomized runs), and performance was evaluated with a quadratic cost that balances tracking error and control effort. The results show that principal-axis control did not outperform the body-frame approach; in many cases, the body frame was favored. As an independent alternative, we also tried Model Reference Adaptive Control (MRAC). In a linearized setting, the implementation tracked the reference model well and canceled matched uncertainties. Future work focuses on modifications to make the controller applicable to nonlinear dynamics.
Specialet er udført i samarbejde med den danske satellitproducent Space Inventor. Målet var at forbedre holdningskontrol, når satellittens masse er ujævnt fordelt. Space Inventor foreslog en koordinattransformation, så styringen kunne ske langs hovedakserne – satellittens naturlige rotationsakser – for at modvirke effekten af masseubalancen. Denne idé blev realiseret med to klassiske regulatorer: en PD-regulator (proportional–derivat) og en LQR (lineær kvadratisk regulator), og sammenlignet med direkte styring i det kropsfaste koordinatsystem. Sammenligningen blev udført med Monte Carlo-simulationer (mange tilfældige kørsler), og ydeevnen blev vurderet med en kvadratisk målfunktion, der afvejer sporingsfejl og styreindsats. Resultatet var, at hovedakse-styring ikke klarede sig bedre end kropsfast styring; i mange tilfælde var kropsrammen endda at foretrække. Derudover blev en alternativ tilgang, modelreference-adaptiv styring (MRAC), afprøvet. Implementeringen viste tilfredsstillende sporing af en referencemodel og kompensation for matchede usikkerheder i den lineære model. Fremtidigt arbejde omfatter ændringer, så regulatoren også kan anvendes til ikke-lineær dynamik.
[This apstract has been rewritten with the help of AI based on the project's original abstract]
Keywords