Physics-informed LLM Agent for Space Operations

NASA's real-time missions demand exceptional technical and cognitive performance from personnel under high stress, which can lead to errors. As missions extend to the Moon, Mars, and deep space, resilient, human-centered decision support systems become critical, especially for enhancing crew autonomy without Earth support. While Large Language Models (LLMs) have advanced commercial operations, their potential in the space domain for human decision support remains largely unexplored. Objectives: Demonstrate a novel digital assistant that utilizes a unique hybrid fault diagnosis architecture that integrates physics-based anomaly detection via Physics Informed Neural Network (PINN) with Retrieval-Augmented Generation (RAG) Large Language Model (LLM). Grounded in Physics and real data, this system will 1) Help human operators detect, interpret, and respond to spacecraft/system anomalies faster, more accurately, and with lower cognitive load. 2) Enhance human operator autonomy in decision making, akin to Mars-class latency missions.

Deliverables: o PLASO Prototype on Azure Platform, o Video of demo with select VIPER Rover Mission Ops Scenarioso Patent Application and Conference/Journal Paper, o Final Report containing experimental Results and discussion of app, licability to Artemis, Mars, and deep spacemissions.

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.