Innovative AI/ML methods for onboard processing of real-time data streams to optimize science data return and power usage

Human exploration and robotic missions require rapid assessment of continuous data that scales beyond the availability of a human-in-the-loop for assessment and requires monitoring by algorithms. This is particularly critical for the detection of rare events, whether they be once-in-mission science sampling opportunities or hazards which must be mitigated. Detection for rare events usually cannot be explicitly programmed, and requires flexible algorithms that can make nuanced decisions across diverse data. During the first year of effort, we developed the COntinous Realtime Detecting Infrastructure (CORGI), a code for simulation and machine learning assessment of continuous data via Convolutional Neural Networks. Using continuous seismic data as a test case, we built an all-in-one infrastructure that is able to modify input data to user-specified resolution, run inference using the EQSpec detection code, and measure hardware and power cost on flight-analogous hardware to assess feasibility for segmented telemetry. For the second year of this project, we want to build upon the developed code and expand it to operate on multiple types of data across different tasks to find the limits of this technology with respect to the kinds of problems it can solve. Specifically, we will test this methodology to varied change detection problems in planetary science, heliophysics, and human exploration. Each selected use case will have several models operating in tandem to output predictions. The multi-discipline expertise at Goddard, which will be reflected in a team composed of members across multiple divisions, is crucial for ensuring a final output product that is useful and flexible across multiple tasks.

This is a mature technology (TRL 7+) — the realistic path in is usually NASA's Technology Transfer Program: licensing an existing NASA patent, or a Space Act Agreement to use NASA facilities/expertise directly. NASA also runs a startup licensing program with no upfront fee for companies formed to commercialize a specific NASA technology.

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.