The Universe

Using Autonomous Onboard Science Agents to Enable Groundbreaking Discoveries

Image from animation of stable orbits of eight SmallSat spacecraft
Animation of stable orbits of eight SmallSat spacecraft (labeled “SC,” with each a different color). The view is centered on the south polar region of Enceladus, an icy ocean moon of Saturn where scientists have observed persistent plume activity that ejects material into space. The video shows that the constellation would provide observations of the entire globe. Wayne Yu, Grant Hecht, NASA/Goddard

A group of researchers under NASA Goddard Space Flight Center’s Distributed Systems Missions initiative are working to enable agile missions that are capable of autonomously interpreting and responding to new science data. They chose a challenging but fascinating planetary target for their design—Enceladus, an icy ocean moon of Saturn that has been deemed one of the most promising targets to search for life in our solar system. The team modeled various trajectories and orbits of eight small spacecraft (SmallSats) to produce the first stable constellation around Enceladus that would provide global science observations. The trajectories of these spacecraft were then paired with mass spectrometry data from Enceladus—analog experiments in a simulated mission scenario. They used their in-development autonomous onboard science agent to analyze data in ‘real-time.’ The agent then uses multiple machine learning models to process and interpret mass spectra, come to a consensus about the interpretations, and then make decisions such as requests for follow-on analysis or prioritized data transmission based on mission goals.

Quick Facts

NASA’s supercomputers allow researchers to "dream big"—try out complex processes and maneuvers that have never been performed in space. Researchers can use the NASA Center for Climate Simulation's supercomputing resources to learn how to make processes computationally "light" enough to be deployed on a spacecraft.

Bethany Theiling,
NASA Goddard Space Flight Center
  • It is critical for NASA to design missions capable of autonomously interpreting new science data collected from distant worlds and performing coordinated analyses using minimal contact with Earth.
  • A team of researchers at NASA’s Goddard Space Flight Center is proving this concept by developing and testing models using the supercomputing resources from the NASA Center for Climate Simulation.
  • The project demonstrates that a stable orbit of multiple SmallSats can be maintained around Enceladus and that seawater chemistry and biosignatures can be identified with high accuracy using remote measurements.
  • This research will be used to design and prepare software architecture for a future mission profile that uses a constellation of spacecraft.

Researchers

  • Bethany Theiling, NASA Goddard Space Flight Center
  • Wayne Yu, NASA Goddard Space Flight Center

More Information