% grade is inconsistent and incorrect with respect to elevation plot

[GoogleCodeExporter]

What steps will reproduce the problem?
1. track the same hill more than once
2.
3.

What is the expected output? What do you see instead?
I computed grade based on known elevation data.
myTracks shows elevation plot correctly, yet it's max grade result does not 
correspond to the independently computed grade of the steepest hill on the 
route.  Also, the max-grade number is not consistent when the same route is 
recorded repeatedly.

What version of MyTracks are you using? On what version of Android? On what
phone?

Running on HTC-Incredible Android 2.2;
myTracks version is 1.0.19

If possible please provide a log.  Either upload here or send to mytracks-
support@google.com.
Detailed instructions can be found here:
http://code.google.com/p/mytracks/wiki/HowToReportErrors

Please provide any additional information here:
http://maps.google.com/maps/ms?msa=0&msid=205856267881636957247.00049417d891bbd1
43840

Original issue reported on code.google.com by tgoh...@gmail.com on 3 Nov 2010 at 2:12

• Merged into: #5

[GoogleCodeExporter]

I am tempted to close this issue. We try to smooth out the grade since the 
elevation data we get back from the location provider is really noisy.  The 
reason that the data is not always the same is that we do throw out the noisest 
data we get back.

Original comment by sandordo...@google.com on 16 Nov 2010 at 5:50

• Added labels: Priority-Low
• Removed labels: Priority-Medium

[GoogleCodeExporter]

Hello,

Another possible approach to computing the grade comes to mind.  Your plot
of elevation profile appears to be very smooth without any obvious anomalies
or deviations from the true grade of the course.  Would it be possible to
compute grade from the same array that is used to produce the elevation
plot?  It appears that adequate smoothing and elimination of anomalies has
been accomplished in the data that is used to generate the plot.

Original comment by tgoh...@gmail.com on 16 Nov 2010 at 3:11

[GoogleCodeExporter]

Ok I'll leave it open to reflect that we could do better at detecting the 
grade.  I think using the elevation plot would be a mistake as that samples too 
aggressively, but we might want to look at computing the grade over locations 
that are at least n meters appart instead of looking at sequential points.

Original comment by sandordo...@google.com on 18 Nov 2010 at 4:12

[GoogleCodeExporter]

Let me know if I can help by sharing any specific data sets with you.  I
have one cycling course that has a known set of highway grades on it; and I
can share multiple passes of this course if that would be of help.

Thanks for the efforts!

-Tom Gohl

Original comment by tgoh...@gmail.com on 18 Nov 2010 at 4:19

[GoogleCodeExporter]

Original comment by sandordo...@google.com on 25 Feb 2011 at 6:08

• Changed state: Duplicate