The cooperative control of satellite clusters in elliptical, low-Earth orbit is studied, with the goal of minimizing the necessary information passed among the individual satellites in the cluster. We investigate two possible control paradigms in this paper. The system is described using linearized equations of motion, allowing it to be expressed as a time-varying linear system. The control objective is to attain a required formation at a specified point along the orbit. A decentralized controller is used, in which each satellite maintains a local estimate of the overall state of the cluster. These estimates, along with any control information, are shared after any satellite executes a control action. The second paradigm is an extension of the first, in which state estimates are never shared, and only the control information is passed. In each case, less information being passed results in a higher computational burden on each satellite. Simulation results show cyclic errors, likely induced by higher-order terms in eccentricity and inclinations. The controller that shares state estimates performs much better than the controller that passes only control information.
