New Decentralized Algorithms for Spacecraft Formation Control Based on a Cyclic Approach
MASSACHUSETTS INST OF TECH CAMBRIDGE DEPT OF AERONAUTICS AND ASTRONAUTICS
Pagination or Media Count:
When considering the formation control problem for large number of spacecraft, the advantages of implementing control approaches with a centralized coordination mechanism can be outpaced by the risks associated with having a primary vital control unit. Additionally, a centralized approach implies an inherent difficulty in gathering and broadcasting information fromto the overall system. Thus, there is a need to explore efficient decentralized control approaches. In this thesis a new approach to spacecraft formation control is formulated by exploring and enhancing the theory of convergence to geometric patterns and exploring the analysis of this approach using the tools of contracting theory. An extensive analysis of the cyclic pursuit dynamics leads to developing control laws useful for spacecraft formation flight which do not track fixed relative trajectories and therefore reduce the global coordination requirements. The proposed approach leads to local control laws that verify global emergent behaviors specified as convergence to a particular manifold. Analysis of such control approach by using tools of partial contraction theory is performed, producing important convergence results. By applying and extending results from the theory of partially contracting systems, an approach to deriving sufficient conditions for convergence is formulated. Results of the implementation of these algorithms were obtained using the SPHERES testbed on board the International Space Station, validating many important properties of this decentralized control approach. To complement the results we also consider a short analysis of the advantages of decentralized versus centralized approach by comparing the optimal performance and the effects of complexity and robustness for different architectures and address the issues of implementing decentralized algorithms in a inherently coupled system like the Electromagnetic Formation Flight.
- Spacecraft Trajectories and Reentry