Conflict Resolution and Optimization of Multiple-Satellite Systems (CROMSAT)

Abstract

This thesis produces models of satellite constellations using finite state automata (FSA) or finite automata (FA) and optimizes the sequence of targets for two missions. Two simplified FSA models of satellite constellations with one ground control station (GCS) are developed. The first model is of a single spacecraft and the second includes two spacecraft. Based upon the language, states, and state transitions of each model, the author transforms the FA into a network and enumerates the shortest paths for indicative lists of meta-tasks from each model. The first model is provisionally implemented in MATLAB. The author finds two separate optimal target selection sequences for randomly generated sample target sets using commercial off-the-shelf optimization software. Although stochastically fabricated, the sample target sets reflect valid scenarios for a satellite imagery mission. The first sequence, a traveling salesman problem, minimizes the time required for processing all targets given a multiple orbit mission. For a representative sample target set, this is 2.34 orbits. The second sequence, a prize collecting traveling salesman problem, maximizes the number of targets processed given a dual orbit mission. For the same sample target set, two orbits permit the processing of seven targets.