Model–Based Reasoning Applied to Biological Spacecraft Payloads for Anomaly Management

With the ever-increasing complexity of spacecraft, the real-time data and state of health analysis by mission operators becomes an intricate process subject to the pitfalls of error-prone human reasoning techniques. If even detected, characterizing an anomalous state on the spacecraft can take substantial amounts of time thus reducing overall operational efficiency and possibly even jeopardizing the mission. This research specifically addresses the state-of-health analysis of biological payloads flown on NASA missions such as GeneSat-1 and PharmaSat. The complex engineering systems and timely anomaly resolution required for maintaining life support in these biological spacecraft makes human diagnosis on its own insufficient. To address these challenges, this project incorporates the use of model-based reasoning for managing anomalies found on board biological microsatellites. A spacecraft payload model was constructed with behaviors relevant to biological sample growth and the associated micro fluidic life support system. A suite of algorithms was applied to the model for computing diagnosis conjectures on detected anomalies. Implemented in MATLAB/ Simulink and designed for use as a ground-based tool for human operators, this system focused primarily on mission operator decision support. The system was developed via analysis of GeneSat-1 flight data and with a biological test bed which emulated the growth characteristics of the spacecraft. It was later integrated into the ground segment of the PharmaSat space system and verified with its flight data after launch. Results gained from experimentation validated the tool’s ability to reason on unanticipated anomalies, its speed of analysis, and its ability to augment a human operator for real-time anomaly characterization and decision support