Kuri localization performance

[nickswalker]

We've had a hard time getting Kuri to map or navigate. Some of this is probably because of the scale of the space (can't expect the robot to stay localized when there's nothing in the ~8ft range of the lidar), but even in Kuri-scale areas, we observe a lot of drift, especially as the robot rotates.

Here's a video showing what I mean.

Any tips for improving localization performance?

[Carter90]

Looks good to me, I did not see any major drift in the video.
I have noticed that my Kuri does get lost in large spaces with lots of moving people, chairs and rolling whiteboards(that the laser cant see) but that is totally expected.
If you are talking about the planner stopping thinking it arrived at its goal when Kuri is still half a box away, that is probably just a parameter for how close to the goal is okay.
Can you provide your Rvis perspective and how you at attached to the planner so I can try to reproduce?