Radar for vital signs - mmWave sensitivity in real rooms?

mmWave vital-sign radar works in demos but real rooms have multipath and movement everywhere. What sensitivity did you achieve in occupied rooms, and what signal processing saved you?

mmWave vital-sign radar demos look clean because the subject sits still in a low-clutter booth. Real rooms add multipath, micro-motion from fans/HVAC, and other people.

Sensitivity that survives a room:
1. Choose the right Doppler / range bin: chest motion is sub-mm to mm at 0.1-2 Hz; clutter from curtains and displays can dwarf that unless you range-gate tightly and high-pass / notch known interferers.
2. Multipath: a strong mirror path can create ghost Doppler. Multi-chirp averaging, angle bins (even a small MIMO), or simply moving the sensor off specular walls helps more than raising TX power.
3. Body orientation and clothing: side-on and heavy fabric cut the RCS of chest motion; fuse with a second view or lower your detection claim to "presence / gross motion" when the geometry is bad.
4. Self-motion: sensor vibration and cable flex look like vital signs - stiff mount and a static clutter map beat DSP heroics.
5. Validate with contact reference (belt / PPG) during bring-up; tune CFAR and filter cutoffs on that, not on a marketing GIF.

Share frequency (60 vs 77 GHz class), chirp BW, and whether you need heart rate or only respiration / presence - that sets how hard the multipath fight is.

— rfforum 编辑部 / Editorial Team