GPS drift - adding offsets manually?

edited August 2020 in General

When comparing driving lines at the same track but from different days the GPS drift will make it hard to see what is going on since they can be placed a few meters differently.

I can make a script that adds or subtracts to lat and long to all data points in an exported log file. But... have you ever considered an offset function in RC for this? For example with arrow buttons moving the drive line step by step till it's more correct on the map?



PS:

Also, buying RC and getting all upgrades over the years seems unfair. Unfair to you. You should have a button in the app for a voluntary "upgrade fee" when releasing massively new things in the app 👍

Comments

  • I'm unsure of this feature... Which GPS were you using BTW? A good GPS with a support for the DGPS satellite _should_ fix the day-to-day spread, but I guess it does not always help.
  • edited August 2020
    I have used Dual XGP160 for a long time but this year I'm using Garmin GLO2

    But even on the same day I get drifts of maybe two meters making the map useless. These laps were within decimeters the same.


    http://magnus.thome.nu/tmp/map.png

    http://magnus.thome.nu/tmp/mapgraph.png
  • I observe the same problem at AutoX. There was just no way for such variance at that course: https://imgur.com/jTf1o3L

    I can share the RaceChrono session file from that event if useful.

  • Here's another example: https://imgur.com/BM30IAS
    Here's an onboard from that lap, you can clearly see I'm driving over the apex curb at 1:55:
  • edited August 2020
    @timurrrr Keep in mind that GPS position is not absolute, and the current GPS receivers we use are not good enough to always hit apex and stay on track. New chipsets are coming with Galileo support, and we should be able to compare driving line reliably in a year or two...

    Looking at your screenshots, the data looks pretty good quality, I think theres only like 5 meters variation on the data if I'm looking it correctly? But of course on a small circuit like that 5 meters looks like a big deal.
  • Yeah, understandably that the civilian version of GPS isn't precise enough. But it seems that in most cases for each AutoX run or a track lap all the positions are just offset by roughly the same amount. If we were able to manually correct for that per run, that would probably allow us to eliminate ~90% of the issues.
  • I'm considering if I should do one coordinate adjustment per session (or session resume), but per lap adjustment is not really possible.
  • Per resume might be a good start
  • Per resume. I tend to do resume when I get back a completely different day sometimes

  • 1: First of all, the positioning data provided by GPS+Glonass is accurate enough. The problem is that the GPS hardware cannot receive good GPS signals, resulting in a decrease in position accuracy. The root cause lies in the location where the GPS antenna is installed. The outer shell of the vehicle blocks most of the GPS signals. When you put the xgps160 on the roof of the car, the accuracy will be improved. The addition of Galileo and Beidou will completely change this problem.
    2: The deviation problem of Google satellite map. I only found this problem in China. Google satellite maps of North America and Europe did not find any deviations.
    3: Manual adjustment is a function that China needs.
Sign In or Register to comment.