Inspiration

About 38,000 deaths are linked to sleep apnea in the US alone. Many devices catering only towards infants leaves adults that are incapable of self-care in the dark. Furthermore these devices are also unreliable as their biggest pitfalls are accuracy, false positives, and worst of all: false negatives leading to preventable deaths and irreparable damage.

What it does

This app tracks breath frequency and depth using a spring-loaded potentiometer hooked into a Raspberry Pi 5 running on QNX OS.

How we built it

The system uses a hardware component to detect respiration and software to relay information quickly:

  1. Breath Detection: A spring-loaded potentiometer allows for a string around the ribcage to expand and retract when an individual breathes. This allows us to track the respiratory data necessary for analysis.
  2. Interpreting Data: The software is run on QNX OS with a C++ script to keep the latency as low as possible. Then sent to the backend server via HTTP.
  3. Backend: The raw data is processed and stored in a PostgreSQL Database. REST API creating through Express using TypeScript for compile time debugging. All deployed through Railway.
  4. Frontend: Displayed on IOS through SwiftUI.

Challenges we ran into

-Lack of MacOS support for QNX -Missing hardware from kits -3d printing issues

What's next for Sleep Apnea Detection

-Implementing machine learning to adjust for each individuals breathing patterns. -Websockets for lower latency

Share this project:

Updates