Application of optimization techniques to the design and maintenance of satellite constellations

Abstract

Optimization techniques were studied and applied to a variety of applications in both the design and maintenance of satellite constellations. Powell's method and parallel genetic algorithms were used in conjunction with precise orbit propagation schemes to develop robust orbit optimization tools. Specifically, local and global optimization methods were used to design a 113:14 repeat ground track variant of the Ellipso TM inclined elliptical sub-constellation and a gear array variant of the Ellipso TM equatorial sub-constellation. The resulting optimal constellation designs were found to maintain stability, even when subjected to full perturbation analysis. The global optimization technique of parallel genetic algorithms was also used to create an optimization approach capable of maintaining the designed orbits over specified lengths of time. Although the global method proved successful over short time periods, limitations of the approach eliminated longer time span optimizations and led to the creation of a more operational station-keeping optimization scheme. The more operational station-keeping implementation yielded similar station-keeping estimates while allowing for the study of longer time periods.