CardioSync
An ESP32 Holter monitor that records 3-lead ECG and ships every session to AWS for filtering, heart-rate detection and reports.
The problem
Holter studies need days of continuous ECG recording, and getting that data to clinicians in a usable form is still a manual step.
How it works
A dedicated FreeRTOS task samples two AD8232 front-ends at 250 Hz, decoupled from SD writes through a queue. When online, the device requests a presigned URL over mutual-TLS MQTT to AWS IoT Core and streams the file straight to S3. Each upload triggers a containerized Lambda that filters the signal, detects R-peaks and publishes plots, a CSV and a JSON summary. Terraform provisions everything and generates the firmware config.
Pipeline
- 01ECG capture
- 02SD storage
- 03MQTT request
- 04Presigned URL
- 05S3 + ECG Lambda
Where it's used
- Ambulatory cardiac monitoring (Holter studies)
- Biomedical engineering research on low-cost wearables
- Remote patient monitoring in places with intermittent connectivity
Highlights
- Jitter-free 250 Hz sampling with vTaskDelayUntil
- Offline first: binary sessions on SD, retried automatically
- One X.509 certificate per device with a least-privilege policy
- Band-pass, notch and wavelet filtering with R-peak detection
Stack
Want something like this for your company?
I build production versions of these systems on your data and your infrastructure.
Contact me