Re: Speed Drop and Spike in Video Overlay
Posted: Thu Jan 08, 2015 4:39 pm
Here's the spreadsheet: https://dl.dropboxusercontent.com/u/915 ... _Lap9.xlsx
Calculating acceleration data from a position time series is problematic. What you're doing is taking a second derivative. Taking a derivative enhances noise. If you filter out all the noise, you're guaranteed to be filtering the signal as well. That's why recalculating acceleration from the GPS data should be a last resort.
I haven't had problems with corresponding with Dual by email, or at least not the total lack of response that some complain about. I had a bad battery in my original XGPS160, it wouldn't hold a charge. I sent it to them after getting the appropriate instructions and permissions and received a new unit promptly. Updating the firmware was problematic for a while, but that seems to have been fixed too.
Something else you might try is to export the lap data in .vbo format and load it into Circuit Tools from Racelogic. The program is free. It also calculates speed and acceleration from the GPS data and has a range of filtering options. I suspect you'll still see the step changes in speed.
http://www.racelogic.co.uk/_downloads/v ... tTools.zip
What I haven't done, because it's too complicated for my rather minimal programming skills, is to modify data externally and then import it back into HLT.
Calculating acceleration data from a position time series is problematic. What you're doing is taking a second derivative. Taking a derivative enhances noise. If you filter out all the noise, you're guaranteed to be filtering the signal as well. That's why recalculating acceleration from the GPS data should be a last resort.
I haven't had problems with corresponding with Dual by email, or at least not the total lack of response that some complain about. I had a bad battery in my original XGPS160, it wouldn't hold a charge. I sent it to them after getting the appropriate instructions and permissions and received a new unit promptly. Updating the firmware was problematic for a while, but that seems to have been fixed too.
Something else you might try is to export the lap data in .vbo format and load it into Circuit Tools from Racelogic. The program is free. It also calculates speed and acceleration from the GPS data and has a range of filtering options. I suspect you'll still see the step changes in speed.
http://www.racelogic.co.uk/_downloads/v ... tTools.zip
What I haven't done, because it's too complicated for my rather minimal programming skills, is to modify data externally and then import it back into HLT.