Saber Astronautics proposes spacecraft subsystem control software which can autonomously reconfigure avionics for best performance during various mission conditions. The approach is to construct performance metrics from spacecraft health and welfare telemetry and learn their relationships in a probabilistic, multi-modal model called a "System Map". In this way the relationships between consumables, mission performance, and spacecraft subsystems can be learned in the same data driven model. The significance of the innovation is the ability for immediate, accurate real time assessment of the spacecraft systems-of-systems performance. Immediate assessment leads to immediate action, so the System Map is useful for any spacecraft reconfiguration task, spacecraft repair task, or mission decision assessment. It can be used in a fully autonomous, decentralized fashion by implanting on an FPGA as investigated in this proposal, or as a decision aid for a human crew.
More »