M. Bengtson, A. Barrie
Aurora Engineering, New Hampshire, United States
Keywords: data fusion, autonomous operations, event-driven operations, spacecraft autonomy, C-UAS
System operators in domains like spacecraft operations and C-UAS face overwhelming data streams and must make high-stakes decisions in real time. Traditional support tools rely on brittle “if-then” rules or opaque machine learning models - neither of which offers the flexibility, trust, or explainability needed in mission-critical settings. MEDOS offers a transformational alternative: a transparent, deterministic, and fully explainable system that encodes SME knowledge into real-time operations. Unlike machine learning, MEDOS requires no large training datasets and can be rapidly adapted to new telemetry points or evolving mission scenarios without retraining. Specifically, MEDOS can draw conclusions at the swarm level, rather than the individual UAS level. MEDOS delivers quantified confidence levels for detected events - such as spacecraft anomalies or coordinated UAS threats - enabling operators to take the appropriate action. It has achieved TRL 7 on a NASA flagship satellite mission and is ready for further infusion as a compiled C library. By combining human-trustable reasoning with real-time decision support, MEDOS bridges the gap between manual oversight and autonomy. MEDOS will reshape operations in aerospace, defense, and critical infrastructure by reducing cognitive burden, increasing operational resilience, and accelerating the safe adoption of autonomy in high-stakes systems.