SAFE-CARE
The Architecture
Five layers: perception, digital twin and risk, constrained reasoning (LLM + Safety Gate), action/HRI, and care-network interfaces.
SAFE-CARE connects perception, a Resident Care Digital Twin with risk models, LLM reasoning behind a deterministic Safety Constraint Gate, action/HRI, and care-network interfaces into one coordinated pipeline.
Paper diagrams and layer detail will land here. Until then, this page orients visitors and points to contact for architecture discussions.
Results shown here are module-level and/or simulation benchmarks. They are not clinical validation. No living-lab study has been performed.
