Is this a normal inaccuracy of the gps ,is there anyway of offsetting, so the racing line is towards the apexs …it’s way too off from the apexs in the picture
Yep. Just what you get sometimes if the quality and/or number of satellites changed
I looked into this a while back when I to tried align with the images on Google Earth. It’s not a simple shift or rotation to align things.
Not even a linear transformation that i could tell. So I gave up.
My assesment was the GPS was OK but the projection of the aerial photo is not to the accuracy we need.
The actual GPS coordinates are shifted. I do all my data processing outside of RaceStudio, and AIM results still get shifted sometimes.
The reason it doesn’t look like you can shift it is because of the Mercatur projection skewing things. You actually can shift to line it up, but it is hard to define what is “right”. It is usually safe to assume you cross the start/finish at the same point within a small margin and use that as a reference, but, if you were offline passing someone one lap, then that can also be way off.
I wrote a tool to automatically shift the lap to another “good” lap using an RMS Error optimization routine, so maybe I’ll play around with that a little more. It can kick out a corrected CSV file that can be imported back into RS3, but that’s not a process you’d want to do every time.
The rule of thumb I’ve seen is that drift “within a session” is minimal but drift “between sessions” can be substantial, both because of the quality of satellite lock and the error in the overlay
That tracks with what I’ve seen. If you start the kart underneath an awning or near a building, the variability is larger than if you start it out in the open.
The worst was one time when the time synch between the sensor clock (rpm, temps, etc) and the GPS clock (location, acceleration, speed) was off by a few tenths. AIM said that happens when the GPS has trouble getting started, so I had to put a check in my routine to look for that and align the clocks.