rsasaki0109/li_slam_ros2

Covariance Matrix for the odometry ?

Opened this issue · 2 comments

From running the li_slam_ros2 in my project, I can see that the covariance matrix for the odometry is set to 0 is their anyway to get calculated values like the original LIO-SAM did ? I would like to add li_slam_ros2 as an odometry source for the robot_localization stack :D

What is the topic name with covariance in LIO-SAM that I can refer to for consideration?

It seems that you can obtain covariance with the marginalCovariance function in ISAM2 of GTSAM.
In g2o, it appears you can output the error covariance for each vertex using the computeMarginals function.