Way out of my price range! :)
I scanned through most of the paper, thanks. It kept crashing at the end.
I didn't notice anywhere that there was much of a focus on stride frequency, though a good job was done matching the accellerators with stride length. I am impressed with that.
> barring calibration issues, the footpod accurately measures the distance of each individual stride.[/quote]
Which brings up the point, IF it does measure stride length accurately, then why is it off by an average 3 percent difference, i.e. a 6 percent variance? That is 24 second at a 6:40 pace. I have run 17 milers and been within 1 second of my planned pace every mile.
The only way that it could be off that much, provided stride length measurement is accurate, is by NOT considering stride frequency. Otherwise the calculation of pace would be quite precise.
Also, the article pointed out that "no calibration is needed", certainly a major mistake.
I think the error could be greatly reduced by connecting the length predictions with actual stride rate, and allowing the user to calibrate the device.
> Did your "considerable analysis" go so far as estimate a magnitude of expected error after calibration, and what sorts of things can affect accuracy?
My analysis consisted of a notebook of hand calculations. I did consider statistical probability at the time.
Yes, I included calculations barefoot vs with shoes, on the flat and up and down hills.
Aren't gps devices more accurate than 3 percent?
Hopefully they will keep improving the devices.