Distributed Multi-Agent System for Autonomous Satellite Telemetry Processing

Intelligent agents offer an exciting opportunity to reengineer existing satellite operations by reducing workload and increasing productivity of ground operators, reducing telemetry processing, and automating command streams. Here, we propose a distributed multiple software agent system for satellite command and telemetry processing to accomplish these objectives. Our multi-agent system consists of: an on-orbit reflexive agent for mapping satellite events into control actions and alerts; a ground-based learning agent for automating operator responses to event sequences and learning command sequences; and an interface agent responsible for monitoring and controlling the on-orbit and the learning agents and interfacing with the operator. The innovation is the incorporated or several synergistic technologies: the distributed processing and communication capabilities, embedded knowledge, inferencing power, and on-line learning and adaptability of software agents; and an architecture for human computer cooperative problem solving for integration of multi-agent systems. The result of our Phase I study will be the design, development, and feasibility analysis of a distributed multi-agent prototype system for spacecraft telemetry and command processing.| Benefits: The proposed technology can be directly applied to commercial satellite systems. Applications of intelligent agents exist for ground-based manufacturing and transportation systems, and consumer goods. A distributed multi-agent system could be developed for automating manufacturing assembly lines. A monitoring agent could monitor assembly line performance and provided fault detection. An interface agent could provide an integrated cooperative environment for operator visualization of plan performance, fault isolation and reconfiguration, and planning and scheduling.|