IMU + Markers kalman filter

asked 2014-12-02 04:07:29 -0500

qbranchmaster gravatar image

updated 2014-12-03 14:14:42 -0500

bvbdort gravatar image

Hi. In my robot I have two types of localization: Intertial and localization based on 2D markers. For my IMU I use this package: . I wrote my own localization based on 2D markers using chili software . Now I'm trying to apply Kalman filter to combine these two measurements: IMU data and a robot's position on a local map (2D markers map). I found robot_pose_ekf package, but I don't know how to use it well. My robot's position is floating when robot isn't moving. I think that it might be caused by noises coming from localization based on 2D markers. Should it float when IMU isn't moving? I don't understand it. Robot_pose_ekf subscribes visual odometry. Can I send on visual odometry topic my position on local/global map? Thanks.

edit retag flag offensive close merge delete


If you have velocity from IMU try to publish 2d Odometry from velocities. ros odom setup

bvbdort gravatar image bvbdort  ( 2014-12-03 14:13:26 -0500 )edit

probably because of IMU, you should output its data to see whether that is drifting significantly enough that it was primary cause of the drift; I calibrated mine using the husky_robot code (here it is for reference)

bzhan gravatar image bzhan  ( 2016-05-17 15:23:37 -0500 )edit