Map drift algorithm

How does the map drift correction algorithm actually work? Usually there are multiple steps in a geo-shift correction algorithm like:

  • shift (move)
  • rotate (orientation)
  • scale (shrink/expand)

Does Yarbo map drift correction handle all 3? The reason I ask is because my yarbo is showing a different orientation compared to charging pad orientation and lines. Also, without other anchor points (besides charging pad), I don’t see how anything more than a simple shift (move) could be possible.

What does your diagnostics look like under the RTK section? Sounds like a bad antenna to me.

This is using DC RTK, not NetRTK.

Yep, looks like the right antenna (which handles direction) has an issue.

How can you tell? Heading dilution of precision (DOP) I assume?

That and your heading should be zero, your L2 should be higher. Single antenna mode is the only reason this is working still.

@Yarbo-Forum the app really needs to tell you when the core is in single antenna mode for an extended period of time and it should also ideally tell you which antenna is blocked.

100%. I’m surprised it was functional at all with those numbers. A permanent condition like that for sure needs to be red flagged. It’s eventually going to loose it way and end up somewhere it shouldn’t be

haha, already wandered off yesterday…

This is post-replacement of us doing right wheel block assembly (motor) so it could have to do with a reassembly error. BTW, asking us customers to do that kind of service work is asking way too much of us.

Kindly, this didn’t answer my original question about map drift…

Since it happened after the track motor replacement I would recheck the internal antenna connections you had to remove for that service. Make sure they’re still tight.

And if it doesn’t get better, swap antennas from side to side to look for change.

Map drift correction, think of it as it moves the piece of paper that is the map around and lines up the docks current location with the cores actual position/direction. So correcting it with a bum directional antenna would not be ideal. You want to see the Yarbo pointing in the right direction, but just sitting off the side of the dock somewhere. When you correct for map drift, it slides the core on the paper over to the dock to realign the map with the currently reality. Does not do shink or expand and I don’t think it does orientation. If you have a map drift that impacts those aspects, I think that requires a remap.

Ok, what you are describing is my first bullet, shift/translate only. That is most likely not robust enough.

They are tight on the antennas themselves. I should check the other ends.

Make sure you didn’t pinch the cables when putting the shell back on.

We checked them out and right antenna cable was loose on the other end.

Working better now?

The orientation appears accurate and between line grass skips are resolved. But the boundary positions are off. It appears the entire map got shifted north as yarbo stays too far from south line and too close or over north line. I did a map shift with yarbo charging on pad. Hmm ..

If you did the drift fix when it was “broken”, you may need to restore your map from the cloud or a saved local one and reapply your settings, work plans, etc.