Public projects 3

Cimini

Project tutorial by Johan_Ha

  • 974 views
  • 0 comments
  • 3 respects

KY-016 Straight Onto UNO

Project tutorial by Johan_Ha

  • 2,498 views
  • 1 comment
  • 8 respects

From KY-039 To Heart Rate

Project tutorial by Johan_Ha

  • 27,722 views
  • 10 comments
  • 38 respects


Respected projects 1

Bluetooth MIDI-Operated Antique Reed Organ

Project in progress by Willem Hillier

  • 7,896 views
  • 3 comments
  • 26 respects

Comments 3

  • Self Playing Slide Whistle 7 months ago

    Only a professional can tell the difference between the original soundtrack and this!

    What a fun project. I think you can actually improve it a lot. Minimize the movement of the servo poking the sound hole. The friction of the slide of the toy whistle is not big, is it? So the movement could be much faster with the help of some gear and perhaps a better stepper motor and driver.

  • From KY-039 To Heart Rate about 1 year ago

    Have a look at this curve:
    https://hackster.imgix.net/uploads/attachments/459527/finest_n6KNB7vgPJ.png?auto=compress%2Cformat&w=740&h=555&fit=max

    It's the filtered curve. After the first two peaks there's a notch in the curve when it goes down after the peak. The notch is somewhere at 1/3 of the way down. If the curve goes up at the notch, it might be recognised as a heart beat. A better filter would even out that notch.
    It's all about filtering the original signal, plotting the filtered signal, refining the filtering algorithm, doing a kind of a pattern recognition on the plotted signal and measuring the time between repeated patterns. How the signal looks like depends on your sensor, your finger tip, your nail lacquer, your electronics, ambient light and whatnot.
    The best of algorithms would be self adjusting. I guess professional heart rate monitors work like that. Usually one is interested in the mean value of the heart rate at a given time, say within 30 seconds. Finding out how much each beat deviates from the mean might also be interesting, as this paper proposes:
    https://www.ncbi.nlm.nih.gov/pubmed/8044922

  • From KY-039 To Heart Rate about 1 year ago

    If you're running my code and you get readings like 51, 53, 50, 49, 100, 97, 48, 51...
    ...it just means that the filter failed to filter out all noise. The value 100 just means the filtered curve didn't just fall between two heart beats. It made a tiny jump in the other direction and the program thought it was a heartbeat, but it was probably just your finger moving a little between the IR emitter and IR receiver.
    My program really needs some algorithm, which would take care of self calibrating.

Add projectSign up / Login