Multi-Agent Anomaly Resolution System (MaARS)

The goal of this effort is to develop the Multi-agent Anomaly Resolution System (MaARS), which will provide NASA with an Automated Fault Detection, Diagnostics and Recovery (AFDD-R) system assisting in maintaining the health of remote habitats during dormancy. Managing remote habitats, with limited earth support and limited communications such as the Artemis mission, will require novel approaches in which NASA maintains a habitat. Furthermore, remote habitats will be subject to extremely harsh operating conditions such as constant exposure to radiation, micrometeorites, lunar dust, and other hazards. To address these challenges MaARS will use machine learning models to provide advanced anomaly detection and resolution capabilities and use automated and robotic systems to recover from an anomaly.TRACLabs and Texas Aamp;M University (TAMU) are well-placed to provide a solution to this topic and address the challenges associated with providing autonomous capabilities for AFDD-R. Both entities have long histories of developing and deploying software tools that make the remote operation of autonomous systems easier and more intuitive. TRACLabs and TAMU have also identified multiple customers both within and outside of NASA that would be interested in using the MaARS system to achieve their objectives.