Autonomous Agent Cognitive Architecture for Spacecraft Operations

VISIMO#39;s NASA Phase II Proposal, Graceful Architecture for Mitigation of System failures (GRAMS), will continue development of a Cognitive Architecture (CA) for decision support for astronauts, primarily focused on human spaceflight missions. GRAMS is engineered for managing spacecraft operations, primarily addressing autonomous systems management for long-duration space missions. GRAMS employs a modular CA designed to detect, adapt to, and recover from both known and unknown spacecraft failures, thereby reducing cognitive load on crew and enhancing mission resilience. Key features include a Risk Identification Algorithm, Alert Generator, Failure Simulator, and Action Recommender. GRAMS operates in two configurations: Simultaneous Monitor/Training mode for monitoring and alerting, and Incident Recovery mode for failure mitigation assistance. GRAMS is not only pivotal for deep space missions but will have terrestrial applications. VISIMOrsquo;s proposed Ramp;D addresses the need for advanced AI solutions to support mission success and safety in autonomous, deep space, and long-duration missions, aligning with NASA#39;s strategic goals. This is critical for NASA#39;s deep space explorations where rapid and accurate response to unforeseen incidents is crucial. Through GRAMS, VISIMO will produce a software prototype of the CA that (1) includes an intuitive, function user interface; (2) includes modular components, configuration scripts, and documentation to assist in the transition to new missions; (3) performance reports of the architecturersquo;s ability to ingest and process live sensor data; and (4) performance reports demonstrating the reliability and accuracy of the cognitive architecture over long duration missions in a space environment. GRAMSrsquo; modular design will allow easy adoption of the platform into new, crewed, deep space missions, as well as operating in additional spaceborne and terrestrial used cases, both human-directed and autonomous.