Optimal design of grid coverage by satellite constellation using genetic algorithm

Abstract

With recent increase of interest in the space community towards the application of low cost, this thesis tries to capture the idea of designing an optimum orbit for area coverage using a constellation of satellites with low cost sensor for limited coverage radius, around one degree field of view. The optimization problem is focused on non-symmetric, circular orbit constellation over three Low Earth Orbit (LEO) satellites. Optimization is done using Genetic Algorithm (GA), which is chosen due to its ability to find a global optimum solution for non-linear problem, which suit this study case. The area of interest in this simulation is represented as a polygon filled with grid points. In this thesis up to five optimization variables is varied from six classical orbit elements. Both single and multi-objectives optimization are done in this thesis. While the single objective optimization only use percent coverage as the objective, the multi-objectives optimization uses percent coverage and revisit time as two main parameter to analyze the performance of the constellation. Some effort are done to improve the objective function value and to minimize the computational time. Semi-analytical approach is implemented in this study for initial guess in the multi-objectives case. This method has proved to be able to reduce the overall computational time. A number of case studies are also done by varying genetic operators; the Selection, Crossover, and Mutation method. To determine the best parametric operator combination, fitness value and the computational time resulted from each study cases are compared.