MAIFLOWER

Modular AI for Faults: Local Online Watch and Efficient Response (MAIFLOWER)

Stottler Henke is developing MAIFLOWER, an AI based system for rapidly detecting and diagnosing faults for the Vertical Solar Array for Lunar Traverse (VOLT) operating on the lunar surface. MAIFLOWER will be integrated into the first VOLT lunar rover test on the moon, which is planned for 2029. MAIFLOWER is NASA funded, and we are collaborating closely with the team at Astrobotic who is designing, building and testing the VOLT lunar rover. The VOLT rover will egress from its lander, transit to a desired location near the lunar South Pole, dig into the lunar soil by oscillating the wheels, and level and deploy a 20 meter tall solar array which will generate power for manned and unmanned systems operating on the lunar surface. MAIFLOWER is a critical capability because human operators on the ground will not be able to respond quickly enough to critical mechanical faults that could jeopardize the mission. The VOLT can experience mechanical faults while navigating to its destination, performing the solar array leveling process, or while the solar array is deployed and tracking the sun. MAIFLOWER will also detect and diagnose complex electrical power system (EPS) faults. The VOLT rover has an internal EPS and is connected to the lunar power grid. The lunar power grid will consist of multiple VOLTs connected to electrical loads including vehicles, habitats, and experiments. Overall, MAIFLOWER will significantly reduce the risk of the VOLT mission.

MAIFLOWER makes use of a suite of Artificial Intelligence (AI) technologies, including both model-based and Machine Learning (ML) algorithms. This hybrid approach leverages the strengths of each fault detection and diagnosis method and allows algorithms that are stronger in certain scenarios to complement methods that are weaker in those cases. Our hybrid fault detection approach is widely applicable to different spacecraft, subsystems and can be integrated with most spacecraft software systems. We have integrated our hybrid fault detection software into the robot operating system (ROS) framework, for the MAIFLOWER project and have integrated with core Flight System (cFS) on other projects.

Stottler, R., Finnigan, E., Ramachandran, S., Singhal, A., & Healy, C. (2025). Modular AI for Faults: Local Watch and Efficient Response. Presented at the 2025 IEEE Aerospace Conference. Big Sky, Montana, March 1-8, 2025. Paper.

Stottler, D., Ramachandran, S., Healy, C., & Singhal, A. (2023). Autonomous, Hybrid Space System Fault and Anomaly Detection, Diagnosis, Root Cause Determination, and Recovery. Proceedings of Advanced Maui Optical and Space Surveillance Technologies Conference (AMOS) 2023. Paper.