Elderly · Strong theory
Gait Decay Early-Warning (Fall Forecasting, not Fall Detection)
Notices walking getting slower and shakier weeks before a fall is likely.
The physics is solid; our own validation is scheduled.
How it works
Walking stirs the WiFi waves in a rhythm — faster steps, longer strides, steadier swing all leave their mark. The system quietly measures walking speed and rhythm every day and watches the trend. A slow decline over weeks is the earliest warning sign of a coming fall.
In the physics: A walking torso produces a bulk Doppler shift proportional to velocity (~10-25 Hz at 2.4GHz for 0.6-1.5 m/s), while swinging limbs add periodic micro-Doppler sidebands at stride cadence. Slowing gait, shortened stride and increased double-support time compress and blur this micro-Doppler signature in physically predictable ways; sit-to-stand shows a characteristic vertical-velocity chirp whose duration lengthens with lower-limb weakness.
See it work
This is real pipeline output for Elderly, replayed. The exact reading depends on your room.
Go deeper
How the field becomes an answer
How we prove it
OPERAnet gait/activity segments for micro-Doppler gait feature extraction; own-hardware longitudinal study in 10 assisted-living units against monthly Timed-Up-and-Go and 4m walk-test ground truth; fall AUC 0.8 already proven (class A) anchors the acute end.
Who it is for
For assisted-living executive directors: 'We tell you WHO will fall NEXT MONTH, not who fell last night' — a per-resident fall-risk trend line with zero wearables and zero cameras.
runtime Runs today in LatentField 2.0 as fall_detect
The matching solution
Complete first-room estimateAging-in-Place Guardian3 nodes · ~$130 parts · 60 min Request kit, subscribe, or set up →Compatible DIY hardware options
15 min · beginner
60 min · intermediate