-
Notifications
You must be signed in to change notification settings - Fork 157
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Extreme Odometry Drift #204
Comments
Just an update, I'm still seeing this issue. I recently did a fresh install from the OS up on a new computer, and I also see this issue on my new computer. |
Ok, another update, I ran just the kimera-VIO example (without the ros part) on MH-01-easy data, and on the same data in the ros version, and compared the output odometry. This is what I got: From that, I can see that the kimera only part seems to be working fine, it doesn't deviate much from the gt data, but the ros data is really far off, so wherever the problem is, I think it's somewhere in the ros stuff. Also noticed that the data runs much, much faster in the kimera only version, not sure if that's a difference between the csv v. bag files, or delays from the ros wrapper stuff. Also the mesh was much denser. |
I'm having the same issue as well. Has anyone solved it? |
Description:
When I run any kimera nodes, the odometry that comes out is way off of what it should be. It drifts a huge amount, enough that I almost suspect it is putting out positions in mm instead of m. Here is a link to a video of the rviz window when I run the euroc example from the main page. I haven't changed any of the config files. I changed the 'use_lcd' parameter to true and added the rviz node to the launch file (rather than in a separate window), and that is it.
Command:
Console output: (only the latest three instances, but these are put out with similar data throughout)
Additional files:
Please attach all the files needed to reproduce the error.
Please give also the following information:
The text was updated successfully, but these errors were encountered: