Cobot Perception HUD — see, decide, act
Monitoring software can only tell you something happened. A SMERT Cobot understands the scene and then does something about it — with an arm, a voice and a mobile base.

Scene nominal · resident seated
Pose stable, upright
Standby — no action required.
Perception runs on the cobot's own edge computer. Frames stay on site; only events, scores and short clips are shared with carers, and safety limits (speed, force, e-stop) remain enforced on the arm itself.
How it works
See
2D cameras plus a SMERT 3D module build a live scene model: people, poses, objects, hazards and distances.
Decide
On-device models score the situation against your care or service rules — no frames leave the site.
Act
The cobot physically responds: clears the hazard, fetches the object, speaks to the person, repositions itself.
Escalate
Only when a human is needed: a short clip, the score and the action already taken go to the carer or manager.