Investigation of Data Augmentation Techniques for Machine Learning Performance on Mass Spectrometry Data

Mass spectrometers are powerful instruments that aim to identify unknown compounds via their molecular weight, as well as perform quantitative analysis, and have been essential to space missions since the 1970s through today (e.g., JUICE, Dragonfly, Europa Clipper). As identified in the Planetary Decadal Survey, missions targeting remote planetary bodies will face limited data rates and volumes and will require onboard science autonomy to optimize science return. In this proposal, we build on our earlier science autonomy work using the Mars Organic Molecule Analyzer (MOMA) instrument for the ExoMars mission as a proof-of-concept. We will use the collected dataset from the MOMA Engineering Test Unit (ETU) at Goddard to apply several machine learning (ML) algorithms for spectral analysis, and this work will contribute to the development of a mature framework for ML tools in mass spectrometry (MS) data analysis, advancing science autonomy for existing and future planetary missions.

We will investigate data augmentation techniques to expand the data volume of our relevant dataset; preparing this methodology now will benefit MOMA operations as well as ongoing science autonomy development for the Dragonfly Mass Spectrometer (DraMS) that is in-development at GSFC (anticipated launch 2028).

Our initial implementation aims at enhancing MOMA scientists’ decision-making process during operations on Mars on the Rosalind Franklin (formerly ExoMars) mission (now planned for launch in 2028, to be confirmed). Our progress in science autonomy will also potentially lead to its online implementation on the Dragonfly mission to Titan, which includes a MOMA-like mass spectrometer. Other future planned and proposed missions to outer planetary bodies would greatly benefit from mass spectrometry autonomy tools due to the slow data downlink to Earth and the limited lifetime of the mission (e.g., 45-day science mission for the Europa lander). Our research in science autonomy will lead to instruments that will make their own analysis of mass spectra and adjust themselves and interact with the spacecraft (e.g., for sampling) in situ to best analyze the sample(s). This would not only allow samples to be analyzed as quickly as they could be acquired, but will also enable the instrument to know which data was decisional and therefore optimize bandwidth by returning to Earth only the most informative or critical science data.

This is early/mid-stage (TRL 3) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.