Run.GPS Online Shop

• Einloggen | Registrieren

• Apps
  • Run.GPS Trainer UV PRO
  • Tägliche Kalorienbilanz
• Downloads
  • Run.GPS Atlas Karten
  • Run.GPS Skins
• Community
  • Ruhmeshalle
  • Top 100 Aktivitätslevel
  • Neueste öffentliche >
    • Trainingseinheiten
    • Routen
    • Benutzer
    • Fotos
    • Trainingspläne
  • Stats >
    • Länder
    • Sportarten
    • Clubs
    • Smartphones
  • Foren
• Routenplaner
• Mein Run.GPS
• Support
  • Foren
  • FAQ
  • Partner

GPS-Sport ForenRun.GPS GeneralHxM distance and speed accuracy disappointing ?

Bitte einloggen, um neue Kommentare zu erstellen

Foren durchsuchen

Erste Seite<Seite2von4>Letzte Seite gmngueko 23.12.2009 13:16:29 UTC yes... yes, like I said before, the HxM sends out the calculated speed, distance, cadence, strides count, heart rate, battery level and some other info, but NOT the 3D acceleration. Lee_Nover 23.12.2009 10:40:06 UTC ahh so the HxM already sends out the calculated speed? well that's no help then :/ I'll send a request for the SDK now .. gmngueko 23.12.2009 01:40:59 UTC HxM data... @Lee_Nover: for me the cadence is the detection of spike in the vertical acceleration after some noise filtering, meaning each time the vertical acceleration is above a certain threshold we declare that there was a step. So from the cadence , for me, you can only tell the frequency at which the vertical acceleration was above a certain threshold, so nothing to do directly with the sum of vector 3 axis components or the vector scalar value. for me The only direct correlation of the scalar value(length of the vector) of acceleration is the speed. Lee_Nover 22.12.2009 19:17:12 UTC HxM data ... from the calculated speed we see in Run.GPS one could assume the reported cadence is the sum of vectors. if that's really the case only they know .. and if it is, then we could do as I said :) tnx for the info, didn't know cadence was exported in csv. I'll check the data soon-ish. gmngueko 22.12.2009 18:54:16 UTC geändert am 22.12.2009 18:59:04 UTC HxM data export @Lee_Nover: RunGPS export in the csv file the speed, distance and cadence, so you have already all the data for analysis actually. the only doubt I have is that usually RunGPS smooth out the GPS speed (average of last x values) but I don't know if it is the same for the speed from HxM, also the minimum value for x is 5. the HxM doesn't send much more which is not exported by RunGPS, the only data missings are (the number os strides and the battery level), the only usefull value being maybe the number of strides (strides count since you started the activity). just to be clear the HxM doesn't send any 3D values, it sends just the already calculated (by the HxM itself with it's algoritm) scalar values for:speed, distance, cadence,strides count. one thing I don't know is : is the the scalar value the amplitude of the vector or is the scalar value the components of the vector in the average moving direction. but those questions should be asked directly to Zephyr, they are the one who knows. Lee_Nover 22.12.2009 18:29:49 UTC Data normalization the received data could be normalized. if we assume that running speed is quite constant with no veering left/right and the up&down jumping (while running/walking) is quite frequent, then we can take the lowest value of the amplitude as the constant movement and use that as the basis for speed calculation. if we go into more detail and depending on the sensitivity of the HxM and the frequency of the data it sends, we could also take into account the acceleration when thrusting forward (when your lifted foot goes forward). one single wave could then be taken apart into two waves (FFT), remove the vertical one and do calculations on the remaining. to analyze the data, you can collect the data from HxM, put the values in excell (column A) and in the next column (B) enter =SIN(B1). select both columns and create a graph "XY Scatter with smooth lines". then you can more easily analyze the collected data :) To Run.GPS team: is the cadence data stored in any of your files? and if so, could it be exported the same as the heart-rate data is? that could help analyze the data we already collected. for now I could extract it manually if the underlying format was revealed :) will attach an image with data sine, current calculated speed (red) and actual speed (green) [1]speed calc gmngueko 22.12.2009 11:20:38 UTC HxM data @Lee_Nover: indeed the HxM data only gives the speed, distance and cadence but don't gives the 3 axis acceleration that leads to those. So the calibration will have to be done on the result speed and distance. the calibration should be feasible as by experience the deviation is quite constant on a given type of movement (walking, running). but it might not be the same calibration for everyone as I see that for instance Tom had some good results while walking, unfortunately he didn't elaborate on his result while running. I'll also point out that this is my second HxM, I sent back, to Zephyr, the previous one for replacement thinking that it was a device fault, but the new one exhibits the same behavior as the old one. if I find times to finalize my small application before the end of my holidays, I can send it to you if you want. @admin: what are your result while running ? is the deviation more or less constant (while walking or running). Lee_Nover 22.12.2009 09:45:58 UTC HxM data no, I don't have any other data. that's too bad to hear :/ guess then there's no right way of fixing this admin 22.12.2009 09:37:44 UTC @Lee_Nover: Yes the HxM does detect data for the 3 axis. But it does not transmit these. It only transmits the absolute (scalar) speed. Or do you have information that I do not have? @dv60ltz: Auf dem Laufband funktioniert die Messung überhaupt nicht, denn da kann keine Beschleunigung gemessen werden, weil man ja relativ zur Erdoberfläche immer an der gleichen Stelle bleibt. Die Trittfrequenz kann aber schon gemessen werden. dv60ltz 21.12.2009 22:18:19 UTC HXM Distanzmessung Hallo Run GPS Team, es ist ja zu diesem Thema schon einiges geschrieben worden. Ich habe auch bei meinen Messungen auf dem Laufband festgestellt, dass beim Laufen die Kilometeranzeige ungenau ist, beim gehen auf dem Laufband wird es dann schon genauer. Eine Kalibrierung ist ja nicht möglich, könnte man vielleicht die Geschwindigkeit regulieren, wie Ihr es mit der Höhenkorrektur macht (+/- Geschwindigkeit)? So könnte man die Geschwindigkeit selbständig verringern oder erhöhen und dem Laufband anpassen, damit müsste sich doch dann auch die Distanz anpassen, oder??? Gruß Olaf Erste Seite<Seite2von4>Letzte Seite