Technology

How the system is put together

Sensing, signal processing, event detection, intervention and notification — and where each part currently stands.

  1. 1

    Sensing

    The belt senses chest and abdominal movement. A foot-worn module adds oxygen saturation and heart rate as a second, independent channel.

    TODAY— Built and demonstrated in the current prototype.
  2. 2

    On-board processing

    The signal is filtered and interpreted on the belt's own microcontroller. There is no separate hub, and no cloud round-trip in the detection path.

    TODAY— Built and demonstrated in the current prototype. Fitting sensing, wireless communication and potential ML inference inside the belt's processing and power budget is an open engineering challenge, not a solved problem.
  3. 3

    Event detection

    The system decides whether a change in the respiratory signal represents a genuine event or a sensor artifact.

    IN DEVELOPMENT— Active engineering or research work. Not finished.
  4. 4

    Intervention

    When an event is detected, the belt applies gentle vibrotactile mechanosensory stimulation.

    TODAY— Built and demonstrated in the current prototype.
  5. 5

    Notification

    The wearable connects wirelessly to a mobile app, which shows the live trace, current vitals and event history.

    TODAY— Built and demonstrated in the current prototype.

Current engineering focus

The prototype works; these are the parts that do not yet work well enough. They are listed as open problems rather than as progress bars, because a percentage would imply a measurement we have not made.

Belt sensitivity

The belt picks up abdominal movement, but sensitivity is not yet where it needs to be for reliable event detection.

IN DEVELOPMENT— Active engineering or research work. Not finished. Stated deliberately as an open engineering problem. Do not soften it into a success.

The processing and power budget

Fitting sensing, wireless communication and eventual on-device inference inside the belt is an open constraint, not a solved one.

TODAY— Built and demonstrated in the current prototype. Fitting sensing, wireless communication and potential ML inference inside the belt's processing and power budget is an open engineering challenge, not a solved problem.

Signal interpretation

Distinguishing a true event from motion artifact is the active research problem, and the reason false alarms are the focus.

IN DEVELOPMENT— Active engineering or research work. Not finished.
Sensing, interpretation, intervention and notification — in the order they actually happen.

The caregiver’s view

The app shows a live trace, current vitals and event history. Its resting state is calm and monochrome on purpose: an interface that looks alarming when nothing is wrong trains people to ignore it.