Bangle.js Heart rate monitor test harness

This repo allows you to test out Espruino/Bangle.js's heart rate monitor code on the desktop.

The actual HRM code is https://github.com/espruino/Espruino/blob/master/libs/misc/heartrate.c

Getting data

• Get a bluetooth ECG heart rate monitor. Suggested: CooSpo HRM808S
• Install https://banglejs.com/apps/?id=bthrm
• Settings -> Apps -> BT HRM
• Scan -> add
• From App Store, go to HRM Accelerometer event recorder : https://banglejs.com/apps/?id=hrmacceven­ts
• Click 'Connect'
• Screen turns red
• Wait...
• Screen turns green and HRM data counter goes up
• Wait at least until 2000 samples recorded. More is ideal
• Click 'Save as CSV', save in data folder
• Edit main.c to add the file to the list

Running

Ensure the Espruino repo (https://github.com/espruino/Espruino) is installed at the same level as this repo, as this code expects to use the code from Espruino itself.

Type make && ./main to run and execute the file (On Linux, Mac or WSL).

Right now, no comparison is done between Bluetooth HRM and calculated readings to get an accuracy value.

CSV data format

The CSV input data contains the following columns:

• Time - Current time (milliseconds since 1970)
• Acc_x,Acc_y,Acc_z - X,Y,Z acceleration in Gs
• HRM_b - BPM figure reported by internal HRM algorithm in Bangle.js (we are running our own algorithm in this repo)
• HRM_c - BPM confidence figure (0..100%) reported by internal HRM algorithm in Bangle.js (we are running our own algorithm in this repo)
• HRM_r - e.raw from the Bangle.on("HRM-raw" event. This is the value that gets passed to the HRM algorithm.
• HRM_f - e.filt from the Bangle.on("HRM-raw" event. This is the filtered value that comes from the Bangle's HRM algorithm and which is used for peak detection (we are running our own algorithm in this repo)
• PPG_r - e.vcPPG from the Bangle.on("HRM-raw" event. This is the PPG value direct from the sensor
• PPG_o - e.vcPPGoffs from the Bangle.on("HRM-raw" event. This is the PPG offset used to map e.vcPPG to e.raw so there are no glitches when the exposure values in the sensor change.
• BTHRM - BPM figure from external Bluetooth HRM device (this is our reference BPM)

Graphs

With gnuplot installed you can run gnuplot -p gnuplot.plot to plot graphs of acceleration vs step count/theshold data for debugging. The plotfiles.sh script will plot all available data as images into the plots folder.

To compile, run and display the most recent graphs, you can do pkill gnuplot;make clean;make && ./main && gnuplot -p gnuplot.plot