Center of mass (in differential drive mobile base) after long time of simulation ( > 1h) stop working ?

asked 2018-01-14 04:45:08 -0500

this post is marked as community wiki

This post is a wiki. Anyone with karma >75 is welcome to improve it.

I'm working on simulation of mobile base (gazebo as a environment for ROS project). During short tests (minutes) everything works as expected. If simulation last longer (> 1h or so) model starts to behave like its center of mass has changed?

If you have any ideas why will be appreciated.

I've attached some pictures and video to show what I am talking about and sources of the model

Mobile base leaning to front even if center of mass is at the back Mobile base front leaning Mobile base front leaning And one more picture with inertias

Strange behavior in video (in the middle of the film)

One more video with strange behavior

edit retag flag offensive close merge delete



Could you post a picture of the robot with inertias enabled?

wentz gravatar imagewentz ( 2018-01-15 06:45:45 -0500 )edit

Sure. I've just edited the question.

wojtas gravatar imagewojtas ( 2018-01-15 10:34:06 -0500 )edit

The inertia of the wheels look rotated by 90 degrees

chapulina gravatar imagechapulina ( 2018-01-15 13:07:52 -0500 )edit

I've just corrected and waiting for results... and results are not good. I've uploaded video in question (One more video with strange behavior - )

wojtas gravatar imagewojtas ( 2018-01-17 13:32:00 -0500 )edit