Traditionally, satellite systems have been composed of stovepiped space and ground components. The payload, spacecraft, ground network operations center, and the mission operations center have been designed with different information systems, thus informa-tion sharing among these disparate systems has been difficult. While such an environ-ment may have been acceptable in the past, today’s mixed-component constellation satel-lite systems are extremely complex and expensive to develop and maintain. Missions are no longer confined to single satellites but span multiple satellites, with the individual sat-ellites each supporting multiple missions. Current information systems cannot handle these missions. A means of integrating the information systems maintained by these satel-lite constellations, and the ability to use this information for the autonomous manage-ment of these systems, is needed. This paper proposes a dynamic, intelligent agent-based architecture that features inter-agent communication, data fusion, data mining, and the ability to resolve semantics over heterogeneous data sources, thus addressing the needs of today’s complex satellite systems.
We apply an agent-based architecture to virtually integrate heterogeneous, complex sys-tems. Our research and development projects are demonstrating that agent-oriented in-formation systems allow information consumers, mission data analysts, and operations personnel to manage ground and space resources, to perform queries relative to any as-pect of the satellite system, to correlate and disseminate mission data, and to mine infor-mation from the heterogeneous data sources that are part of the satellite system. These innovative, agent-based systems show great promise in reducing life-cycle costs, because they automate many of the functions currently performed manually and they enable a smaller staff of less experienced, less expensive personnel to manage complex systems.