The Universe

Toward Science Autonomy for Future NASA Missions

Image from Dragonfly mission preview video, showing the drone approaching the surface of Titan from upper left
Space exploration missions are targeting various bodies in our solar system—including planets, moons, asteroids, and comets—to learn more about our universe and potential habitability beyond Earth. Dragonfly is a NASA mission to explore the chemistry and habitability of Saturn’s largest moon, Titan. The autonomously-operated rotorcraft will visit dozens of sites on Titan with a neutron spectrometer, a drill system, and a mass spectrometer, allowing scientists to make a detailed survey of Titan’s chemical makeup. NASA’S Scientific Visualization Studio, NASA Goddard Space Flight Center

With limited transfer rates and short communication windows, future planetary exploration missions investigating habitability and potential life on distant bodies (e.g., Titan and Enceladus, two of Saturn’s moons) will face communication limitations. Current NASA surface mission operations at Mars rely on ground-in-the-loop interactions that are not suitable for such remote targets. This project focuses on implementing science autonomy tools for mass spectrometry analyses for the Mars Organic Molecule Analyzer (MOMA) instrument onboard the ExoMars mission, the DraMS instrument onboard Dragonfly (Titan), and other mass spectrometer instruments targeting Ocean Worlds (Europa, Enceladus). This research leverages machine learning (ML) and data science to enable science autonomy onboard space missions. The team leveraged the NASA Center for Climate Simulation's supercomputing resources to develop ML algorithms on mass spectrometry (MS) data to support scientists’ decision-making processes, make science operations faster and more efficient, and ultimately to maximize the scientific returns of NASA missions.

Quick Facts

Science autonomy will be required onboard future space missions to enhance and enable space exploration. We use NASA supercomputing resources to develop data science and machine learning algorithms to support predictions of chemical composition using mass spectrometry data for future planetary missions.

Victoria Da Poian,
NASA Goddard Space Flight Center
  • The NASA Goddard team processed tens of thousands of mass spectra data collected over ten years from the flight-like model of the MOMA instrument and then developed ML algorithms—unsupervised visualization algorithms and supervised neural networks (NNs)—that ran on a 20-core Xeon CPU on NASA’s Discover supercomputer over a few hours.
  • They developed NASA’s first science autonomy tool to support mass spectrometry analyses for Mars mission for the MOMA instrument planned to be onboard the ExoMars Rosalind Franklin rover mission (expected launch in 2028).
  • NASA plans to support the ExoMars mission’s MOMA scientists with data science algorithms and NNs running on Earth using flight-like instrument data during future operations on Mars.
  • NASA’s longer-term plan is to use these science autonomy tools (e.g., NNs predictions) onboard a mission to optimize the operations and minimize ground-in-the-loop interactions while maximizing the science return. One of the team's target applications is NASA’s Dragonfly mission to Titan in the 2030s.

Researchers

  • Victoria Da Poian, NASA Goddard Space Flight Center
  • Eric Lyness, NASA Goddard Space Flight Center

More Information