A wearable has to be worn and charged. Short-wave radar has to be close, and works best on someone lying still and uncovered. An under-mattress pad has to carry the patient’s weight through the right mattress. This page sets out what each technology needs, where it does better than ours, and where it does not, using the regulators’ and manufacturers’ own documents.
Every reading depends on the device being on the body with charge left. A flat battery produces no reading and, on most charts, no warning either. A sensor on the wall has no battery to run down.
In a hypertension programme of 6,595 patients, people transmitted data on about 70 per cent of enrolled days. In a Medicaid diabetes programme, fewer than half transmitted on 80 per cent of days.
Smith et al., Healthcare, 2024 · Park et al., Journal of Medical Internet Research, 2023
A wrist or finger device estimates breathing from small changes in the pulse. The XK300 reads the chest rising and keeps the breathing waveform. We could not find a published validation of ring breathing rate against a clinical reference.
Why it is different from a wearable: breathing rhythm and Heartbeat Clarity, explained. Heartbeat Clarity is investigational.
Anything away from the bed: daytime heart rate, heart-rate variability, activity, skin temperature, sleep staging, and a short acute episode where a five-day patch is the right tool. The XK300 does not measure HRV and does not stage sleep.
Weeks and months with no action from the patient: breathing measured from the chest, resting heart rate with nothing worn, bed presence and motion, and a gap in data that is visible the same night rather than thirty days later.
The full ring comparison, measurement by measurement, is on the science page.
Contactless vital-sign radar has been cleared at several frequencies. The higher the frequency, the shorter the wave. The XK300 works at about 7.3 GHz. Two common alternatives work at about 60 GHz and at about 122 GHz.
Breathing has to be read from the chest. At 122 GHz the wave does not get through a duvet; at 60 GHz it is degraded. At 7.3 GHz it passes through the way Wi-Fi passes through a wall.
The shorter the wave, the more a small change in what it reflects from matters. When a patient rolls onto their side, a millimetre-wave return can change completely and the reading has to be found again. A 41 mm wave changes far less.
Where the sensor has to sit behind a protective enclosure or a ceiling tile, a longer wave carries through material that stops a millimetre wave.
“… should be positioned 1.0 meter (3.0 feet) or less”
A cleared 60 GHz device’s FDA summary, on placement to the side of the patient“… RR and HR while the patient is in its detection range at rest.”
A cleared 122 GHz device’s FDA summary“XK300 monitors presence or absence of a patient in a detection area of within 7 meters.”
XK300 FDA summary, K202464. It reads patients on their back or their side.Resting matters to every contactless sensor, ours included: the XK300 reads heart rate and breathing when the person has been still. The difference is how close, and how uncovered, the patient has to be.
Very fine resolution at close range, and small, low-cost chips. For a sensor on a bedside table pointed at an uncovered chest, a millimetre wave can do very well.
Distance, bedding and posture: a detection area of up to 7 metres, readings through a duvet, and placement on the wall, the ceiling, behind a panel or under the mattress.
Each heartbeat and breath gives the body a tiny push. An under-mattress pad feels that push after it has travelled through the patient, the sheet, any topper and the mattress. Counting beats and breaths survives that journey well. Everything the mattress does to the push, it also does to the reading.
The signal is a force that has passed through every layer. The mattress type, its thickness, a second person and the patient’s weight all change it.
The signal is the chest wall’s own motion, read by a radio pulse. Nothing in the sensor carries the patient’s weight.
A US Veterans Affairs assessment reported that an under-mattress system tended to “double read” heart rate from air-mattress vibration, and could not read patients on air mattresses.
VA technology assessment, 2016The same assessment reported that, in some instances, vital signs could not be read for patients positioned on their sides.
VA technology assessment, 2016One manufacturer’s instructions say not to use its pad with waterbeds, air beds or mechanically articulated mattresses; its support pages describe testing on mattresses 10 to 40 cm thick.
Manufacturer instructions for use and support documentationFDA documents for under-mattress systems describe accuracy studied in adults under 111 kg, and an indication for adults up to 350 lb (159 kg). The XK300 has no weight limit, and patient weight does not affect its reading.
FDA 510(k) summariesAt least one manufacturer tells users to ignore the data when someone other than the main user lies on the pad.
Manufacturer instructions for useAn air mat has to lie under the chest. One system’s staff manual explains that if the mat is too far towards the feet, a resident in bed can show no heart or breathing reading and be marked as out of bed.
Manufacturer staff manualSome under-mattress systems are sold as care aids. One maker’s brochure states its product is not designed or made for use in life-threatening situations.
Manufacturer brochure, 2026A care home presenting its experience reported sensors sounding frequently because of sensitivity settings, and a fall that was worked out from the graphs the next day.
Care-home conference presentation, JapanA hospital under-mattress sensor was documented as needing replacement every year; a consumer air pad lists an expected product life of three years.
NICE medtech innovation briefing, 2016 · VA 2016 · manufacturer instructionsLow cost at consumer scale, very simple to set up at home, no view of the room, and good published breathing-rate accuracy in older adults on a suitable mattress.
No load-bearing sensing part. Reads a patient on their back or side, at any weight, through a foam mattress of any thickness, and can move to the wall or ceiling when nothing may go under the bed.
Radio does not pass through water or coil springs. On a waterbed or an innerspring mattress, the unit goes on top of the mattress beneath a foam topper, or on the wall. Hospital beds mostly use foam mattresses; some homes will need the topper or the wall.
No study has yet compared radar and under-mattress heart-rate accuracy head to head in hospital patients, and none has followed either technology’s accuracy over months. We do not claim otherwise.
| Wearables | 60 GHz radar | 122 GHz radar | Under-mattress pads | XK300 · 7.3 GHz | |
|---|---|---|---|---|---|
| Anything worn or charged | Yes | No | No | No | No |
| How breathing is obtained | Estimated from the pulse | Chest motion | Chest motion | Force through the mattress | Chest motion, waveform kept |
| Distance stated in a filing | On the body | 1.0 m or less, to the side | In range, at rest | Under the patient | Detection area within 7 m |
| Posture | Any | Sensitive to position changes | Very sensitive to position changes | Side-lying reported as a problem | Reads on the back or side |
| Bedding | Not relevant | Degraded through a duvet | Does not pass through a duvet | Part of the signal path | Reads through bedding |
| Mattress type | Not relevant | Not relevant | Not relevant | Exclusions stated, e.g. air, water, articulated; tested thickness range | Any thickness; not through water or coil springs |
| Weight limit in labelling | Sizing | — | — | Under 111 kg studied; up to 350 lb indicated | None; weight does not affect the reading |
| Part that carries load or wears | Battery, strap, adhesive | None | None | Pad or sensor; replacement documented | None |
| Away from the bed | Yes, all day | Close range only | Close range only | No | Room presence and motion; vital signs at rest |
“—” means the documents we reviewed do not state it. Entries describe technology categories from regulators’ and manufacturers’ published documents, not any single product, and products within a category differ. Ask us for the full references and 510(k) numbers.