QuickSAT/SHERLOCK, an AI Architecture for Vehicle Health Management, Fault Detection and Fault Management and Complex Human-Machine Systems

sci_Zone, Inc is conducting research to develop QuickSAT/SHERLOCK, a system for Vehicle Health Management, Fault Management and Fault Detection with Fault Classification Functions for deep space satellites, human-rated space vehicles, vehicles operating other terrestrial surfaces and small satellites including those in support of In-Orbit Servicing, Assembly and Manufacturing (ISAM) operations and other mission goals. The QuickSAT/Vehicle Management System (VMS) is a flight proven flight management, Vehicle Health Management plus Fault Detection system flown on DARPA and U.S. Space Force flights and scheduled for flights in 2029. QuickSAT/VMS provides integration also with NASA’s Core Flight System (cFS), Figure 1, and includes the QuickSAT/Xen Space Hypervisor which implements the ARINC 653 partitioning standard, isolating applications running on the same processor. QuickSAT/Xen provides a cybersecurity framework protecting the onboard virtual machines and wireless functions. New to QuickSAT is the integration of the SHERLOCK architecture with our QuickSAT/Autonomy-Grandview environment which integrates an advanced sensing suite, edge-capable Artificial Intelligence (AI), and vehicle tracking tied with vehicle fault management functions to increase the reliability of orbital platforms, provide rapid fault determination, and perceive potential faults that might occur. Generating a useful diagnostic is predicated on having a wealth of insight-rich data to draw on. QuickSAT/SHERLOCK can tie into an on-edge sensing suite including visible light, infra-red, UHF/wide-band radio, plus on-board fault monitoring systems and experiments, and vehicle health data building an internal knowledge base for event and fault classification. This data is fed into a preliminary analysis system designed to extract key features from data. This takes the form of anomaly detection or binary/multi-class status checks (i.e. are the satellite’s thermal management functions operational, experiment state). By collecting many these minor insights, a diagnostic can begin connecting anomalies to gain a clearer picture of any issues at play. SHERLOCK relies on a system built on the Security Risk Taxonomy for Commercial Space Missions authored by Dr. Falco. By comparing symptoms evident in the data analysis with a historical database of security risks, the diagnostic can produce a set of likely candidates for mission failure and propose appropriate action. This component can feed directly into the decision support module of the incident-response architecture and QuickSAT/VMS. The paper will present a comprehensive comparative assessment, provide a detailed concept for autonomy technology to support Gateway operations including visiting vehicles and experiments and satellites in support of ISAM operations, and present a trade study of recommended supporting hardware, and a baseline demonstration of QuickSAT/SHERLOCK.