Autonomous learning agent system for spacecraft telemetry processing

Automation of spacecraft operations can provide a substantial reduction in ground support and program costs. Here, we develop an autonomous spacecraft telemetry processing system consisting of an off-line architecture for the generation and analysis of learned patterns in the telemetry stream and an online mechanism for automated event and alert notification. Our system draws on the embedded knowledge, inferencing power, and learning and adaptability of intelligent software agents.