What the Technology Cannot Sense

Detection features are the reason devices cost what they do, and they are the part least accurately described in any brochure.

When what the technology cannot sense also depends on reliable work records, explore the website can support time, attendance and workload review without being treated as the emergency response itself. Compare any workflow with HSE lone-working guidance and keep alarm ownership, escalation and dispatch responsibilities explicit.

Fall detection, specifically

An accelerometer looking for an impact profile followed by stillness.

Which finds a fall from standing onto a hard floor reliably.

And misses a slide down a wall, a collapse into a chair, a fall onto something soft, and any incapacity that does not involve falling at all.

Sensitivity is a trade. Turn it up and the false alarm rate makes the system unusable; turn it down and it detects less. There is no setting that does both, and a supplier implying otherwise is describing a demonstration rather than a deployment.

No-movement detection

Triggers on stillness beyond a threshold.

Which catches cases fall detection misses, and triggers on lunch, on driving, on paperwork, and on anybody who sits still.

Useful in specific settings — a patrol, a delivery round, work that should involve continuous movement.

Unusable where stillness is normal, which is most desk-adjacent and client-facing work.

Match it to the work rather than enabling it by default, which is a decision most deployments never make consciously.

Automatic against manual

The case for automatic detection is the worker who cannot press anything.

The case against is everything above.

Which resolves into: use it where that specific case is foreseeable and severe, and rely on check-ins elsewhere.

A worker who can press a button is better served by a shorter check-in interval than by detection, because a check-in has no false positives and needs no tuning.

Battery, signal and the physical world

The three failures no feature list mentions.

A device at eight percent at the end of a shift, which is when people are alone.

A basement with no signal.

A unit left in a coat on a hook.

Each defeats every capability above, and each is addressed by habit and arrangement rather than by specification.

What no system can know

Whether the worker is in difficulty but able to act normally, which is the duress case and the reason a code word exists.

Whether stillness is rest or unconsciousness.

Whether an impact was a fall or a dropped bag.

Whether silence means no coverage or a destroyed device.

In each case the resolution is a human decision with context, which is why the monitoring arrangement matters more than the sensing.

Designing around the limits

Layer rather than rely: a check-in interval, a manual alarm, and detection as an addition.

Set the interval assuming detection will fail, because for the cases it cannot see, it will.

And tell workers what it does and does not detect.

A worker who believes the device will notice a collapse may take a risk they would not otherwise take, which is the practical harm of overclaiming and it operates at the level of the individual rather than the organisation.

The question to ask a supplier

"Show me the cases your detection misses."

A supplier who can answer has tested it and is describing a real product.

One who says it detects everything has not, and the gap will be found by your workers rather than in their laboratory.

What to write in the assessment

Which detection is enabled, on what setting, and what it is expected to catch.

What it will not catch.

And what covers those cases instead — which is usually the check-in interval.

An assessment recording "fall detection fitted" without the second and third parts has recorded a feature rather than a control.

Wearables and body position

Detection depends on where the device sits.

A pendant swinging on a lanyard reads differently from a unit clipped at the waist, and manufacturers tune for one position.

Which means a device worn somewhere other than its design position detects worse and false-alarms more, and neither shows up as an obvious fault.

Confirm the intended wearing position and check what people actually do, because comfort wins over instructions within a fortnight.

Environmental interference

Vehicles produce vibration profiles that resemble impacts.

Industrial equipment produces sustained vibration that suppresses no-movement detection.

Cold reduces battery capacity sharply, which matters for outdoor winter work more than any specification suggests.

Test in the conditions the work happens in, not in an office in March.