Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[albert]

Hello Geoffrey,

the 10 sec is OK because you update every 5 sec the calculations, I hope the cpu will allow us to do this and monitor the other things, without getting S&%$#@$.

Good to do AVG is more stable.
Thanks, and finally go and celebrate the new year.


Best regards,

Albert   

[pepelento]

Very interesting topic. I'll try to take some time and think about this. As far as I understand, the problem is the wrong values that on gets (as always). Can you please tell me which kind of values do you have in, lets say, 60 seconds? I think it is important to understand which kind of wrong values do you obtain in the readings and how often they appear. I have some background in Maths, maybe I could help...

[TWG_TECH8]

Hello Hubert,

2. I wonder, if you really have knolls higher than the GPS error, where you are living 
    (okay - sorry - this was a bad joke - but I like bad jokes...)

You will be wondering Ha,Ha, we have 13 km from my place a hill 65 meters high and 1600 meters long, the name , and don't laugh , Amerongse Berg (mountain) . So you see it is not flat, but serious we do go a lot to Germany and Belgium to cycle. Now you know why slope is important to us flat  riders, we than can say E.g. we did 2000 height meters and the steepest hill was 15%, makes us feel good . But when you climb you want to know the percentage if it feels heavy .

Hello Albert,

It sounds challenge and interesting.
Next time maybe you can share your log file and we can do simulation.

Sincerely,

Geoffrey

[AndersBrontosaurus]

Ok. So this isn't an algorithm I'm offering but in the same area...or maybe this should be posted in software...
Anyway, would it be possible, instead of working with algorithm improving the climbing/descending in the computer? Alberts bikeroutetoaster, runningahead and other sites offer routeplanning with accurate heightcharts. How about integrating that in sportstraining gym and make the software adjust the trackpoints downloaded from the watch to more accurate ones?

BR
Anders

[Dardard]

Maybe it should complicate a little, but it should be some interest in taking the HDOP parameter into account in order to validate or compensating the altitude given by the GPS.
Basically, depending on HDOP you could start with something like :
HDOP< 2 period = 10s
2 < HDOP < 5 period = 15s (less accuracy => more averaging)
5<HDOP <10  => periode =20s
HDOP> 20 no measure

Or anything in the same way. The more HDOP is worst is the quality of the altitude measure, so averaging period must raise. The less HDOP is the more accurate you are.

Jean-Yves

[albert]

Maybe it should complicate a little, but it should be some interest in taking the HDOP parameter into account in order to validate or compensating the altitude given by the GPS.
Basically, depending on HDOP you could start with something like :
HDOP< 2 period = 10s
2 < HDOP < 5 period = 15s (less accuracy => more averaging)
5<HDOP <10  => periode =20s
HDOP> 20 no measure

Or anything in the same way. The more HDOP is worst is the quality of the altitude measure, so averaging period must raise. The less HDOP is the more accurate you are.

Jean-Yves
[\quote]


Hello Jean Yves,

This looks the right way, see also the link here. I think so long then HDOP < 2.5 the quality is getting poor.   

Regards,

Albert
http://users.erols.com/dlwilson/gpshdop.htm

[TWG_TECH8]

Hi

The idea of daymaic range control looks ideal case.
We found the HDOP doesn't match the accuracy of location at instant. 


Sincerely,

Geoffrey