![]() | Only 14 pages are availabe for public view |
Abstract Spacecraft Attitude Determination from vectors observations, star detection, centroiding and stars pattern matching are studied in this thesis. Algorithms for star pattern matching are analyzed and a novel algorithm is suggested. It is shown that it supersedes the classical pattern matching techniques like Grid and Triangle matching, in terms of performance and execution times. It is found that the conclusion preferring Grid Algorithms over Graph based ones is not totally correct. In fact, four stars matching in graph based technique outperforms their grid counterpart. It is found that carefully selecting search technique and avoiding pivoting can substantially decrease the algorithm execution time without affecting its performance. A modification to the eigenproblem solution is implemented for the special case of Hermetian Operators. The use of the spectral theorem for Hermetian operator expansion as well as the classical power method enabled us of calculating all the eigenvalues and eigenvectors of a hermitian matrix. Moreover, a complete design and implementation of the interface hardware and software is realized. A framegrabber with programmable logic and its software driver were fully and successfully designed and tested with the camera. Pattern matching algorithms are also tested onboard the star tracker and is found to perform adequately in an acceptable execution time. Software components for attitude determination are implemented based on a native matrix library that we also developed. |