Site Calibration
Establish baselines for beds, chairs, bathrooms and reflective fixtures to reduce false positives.
Care AIoT · Safety Detection
Long-Term Care
Safety Detection System
Thermal, RGB, mmWave radar and edge AI working together — helping residential care facilities and nursing stations stay ahead of fall risk while respecting privacy by design.
System Design
Edge sensing,
central decisions.
Local processing keeps latency low. Detected events are forwarded to the monitoring center or nursing station — fast on-site response, with privacy preserved.
Privacy-Sensitive BathroomsmmWave radar reduces light, steam and privacy constraints.
Living AreasThermal + RGB fusion improves posture and heat-source detection.
Edge AI InferenceLocal decisions on edge devices such as NVIDIA Orin Nano.
Privacy by DesignData minimization, masking, event codes and encrypted transport.
EDGE AI · Real-time
Edge AI Flow
From sensor to alert in five steps
Frames are pre-processed and inferred on-device; only minimized event data is forwarded.
01Image & sensor capture
02Pre-processing & alignment
03RGB / Thermal AI inference
04Privacy masking & de-identification
05Fall analysis & nursing-station alert
Deployment Focus
Start with a focused small-scale PoC
Care environments vary widely. Begin with a target room type — resident room, bathroom, common area, or nursing-station workflow — then tune fall-detection sensitivity from there.
Device Nodes
Each node carries heartbeat monitoring — alerts on offline status or thermal anomalies.
Nursing-Station Alerts
Forward only event codes, zone IDs and minimal thumbnails — no unnecessary raw video retention.
Care Solution