Design and control of microsatellite clusters for tracking missions

Abstract

Space-based tracking missions are an emerging interest that could be accomplished using a cluster of microsatellites. This thesis addresses the design of microsatellite clusters to accurately track a target in a probabilistic suborbital occupancy corridor by pursuing the following: orbit determination using optimal measurement principles, cluster design heuristics and fuel optimal cluster maintenance. These are all evaluated on a high-fidelity simulation testbed. First, the orbital determination approach utilizes optimal measurement principles to design a constellation of clusters that minimizes the average model-based target tracking error. A two part approach, (1) constellation design and (2) cluster design, reduces the overall orbit determination complexity. The constellation design provides continuous, 24 hour coverage of the occupancy corridor and virtual formation centers about which the cluster design formulates the relative microsatellite orbits. Results suggest that satellite separations, rather than the number of the microsatellites in the cluster, are more important for providing target tracking accuracy.

(cont.) Results also show that the J2-induced relative drift of the satellites in a cluster can be reduced by several orders of magnitude with very little degradation in the cluster's tracking capability. Second, this research formulates a cluster design heuristic that provides a robust cluster viewing geometry for a target in any direction. This robust design heuristic provides tracking capability for a cluster that is demonstrated to be comparable to one specifically tuned for a particular target orbit. Third, this thesis presents a receding horizon Model Predictive Control approach to cluster maintenance that exhibits reduced cluster-wide fuel expenditure by allowing relative satellite drift while maintaining mission driven cluster characteristics. The controller achieves this performance by being robust to unmodeled dynamics and noise. Finally, the performance of the integrated cluster-based orbit determination, tracking and control laws is demonstrated on a high-fidelity, multi-satellite simulation testbed. Results include tracking performance and trade-offs as a function of various control objectives.