Howdy,
Thanks to a comment by YT2095 on my last post, I started thinking that I need to do more averaging to achieve a smoother graph. It turns out that this is very simple to do as the Arduino sketch writes 30 BPM samples per line to the text file on the SD card. I created a new script named parse-avg.pl to take the average of every 30 sample set. This created a graph with less fluctuation:
This new averaged data is feed into filter.pl, and a much smoother graph is generated:
Here is a comparison with both on a the same plot:
So I think the plot of the filtered data now is much easier to interpret. However, I need to verify that the data is actually still accurate. I have basic Polar wrist watch receiver which I can use to benchmark. It doesn't log BPM readings, so I will have to write down the reading 30 sec for a short test run (say 10 minutes). Then I can graph the written-down wristwatch readings versus the hrmshield filtered data.
Cheers,
Drew