-
Notifications
You must be signed in to change notification settings - Fork 117
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
Various bug fixes #177
Various bug fixes #177
Conversation
@AndyZe I am sorry for digging this merged solution to ros-controls/gz_ros2_control#113 out, but I think the sagging issue is still there in gazebo classic after this back-ported fix #177. In POSITION control method my robot sags under gravity, and even worse, its joints slowly dislocate ! I built all kuka_jtc_not_maintaining_pos.mp4I was definitively sending a static command (with RQT JTC) to a JTC with position interface but the robot sags down. So its not the "robot not holding position" issue that is another recent issue of JTC it seems. It reminded me of a well known problem from ROS and first use of gazebo 10 years ago to "enforce" a position that by-passes physics (= setting a pose or joint angle and not its effort). You might think, this is a bug, because at the next cycle the position is set back to the correct command, and I would agree, but it seems since gazebo 2 this was never fixed or is a "feature". I don't know. In any case, the solution we used in the past and which also works here to sagging in position mode (without a PID controller converting it to a velocity) is to also set velocity to zero. More generally, I am not sure how this back-ported fix #177 helped with regard to sagging robots. I understand it in gz sim because there seems to be a POSITION_PID implemented and some velocity could be overwritten, but not in gz_classic version of the plugin which has no such thing (yet). old [EDIT] I know that If some algorithm should be tested on a robot sim that has physics interaction (and not only rendering and sensor measurement), then the user should use other control modes than position control of a robot anyway isn't it ? [EDIT2] At least in gz_classic, the sagging is fixed for me. Are you interested by a PR ? |
That being said at least we could have merged this wiy a better name on this repo but oh well :D |
Yes, please add the PR. This might fix UR simulation too |
It does indeed fix UR sim. PR created |
Am I missing something, or is the parameter |
yes, it was added as a parameter, but it's not used in the code |
Should we then remove it again, or will it be used? ;) |
Just a straightforward port of this PR from
gz_ros2_control
:ros-controls/gz_ros2_control#114