-
Notifications
You must be signed in to change notification settings - Fork 1k
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
UR5KinematicsPlugin fails to plan in Cartesian space #252
Comments
@gavanderhoorn should I have just posted here? |
I just install trac-ik-kinematics-plugin and changed my kinematics.yaml file from
to
This fixed both the prefix error and has successfully solved every IK problem I have given it. So, the issue may lie in the planner. Consequently, my issue is solved but the overall issue probably isn't. |
Tagging this as a possible issue to investigate on WRID @ipa-nhg. |
It looks like this was more or less successfully addressed in husky/husky/issues/32. I took a stab at replicating this while using the UR5 demo launch file. I tweaked the provided script to use the planning groups set up by the launch file, and to correct some important omitted parts like the package imports and the node initialization. One problem is that the target point is outside the workspace of the UR5, which is why planning failed when I ran this code. However, I think that the source of the problem @CMobley7 was trying to solve was caused by some behind-the-scenes configuration and parameter setup which was done correctly in
Running the script "as-is":
moveit_commander script output:
Output from RVIZ console:
Running the script after changing the target pose:
moveit_commander script output:
Output from Rviz console:
|
With the extensive analysis by @schornakj I believe we can close this as something too localised. The plugin itself should be OK (to the extent needed for the use-case described), but local configuration seems to have interfered with its operation. If it turns out this is still a problem for the OP or other users, we can re-open. |
To Whom It May Concern,
I ran across issues #127, #155, #184, #202, and #241 in which portions of ur_kimematic/UR5KinematicsPlugin were rewritten to allow the ur5_arm to better plan and execute.
However, the main topic was RVIZ. RVIZ to the best of my knowledge is using forward kinematics due to the fact it already knows the ending joint angles. I attempted to perform inverse kinematic planning using Moveit!’s Python interface with the code below on a Clearpath Husky equiped with a UR5.
However, the code gives the following errors:
The prefix errors do not seem to affect the planner in solving forward kinematic motion problem. As I have used similar code to go to group states set in the srdf, as well as other random valid joint configurations. However, when specifying a Cartesian point, the planner always fails. After reading the posts regarding ur_kimematic/UR5KinematicsPlugin, I think the problem is more likely that my code is not using the same configuration in regards to using limited joint angles when planning and not setting up the same namespace as husky_ur5_planning_execution.launch rather than issues with the planner itself.
Should setting up a launch file like https://github.com/husky/husky/blob/indigo-devel/husky_ur5_moveit_config/launch/moveit_rviz.launch solve the issue. If so, are all the lines necessary and I just need to replace the RVIZ node with my appropriate node? Or is the issue more likely caused by the planner itself?
Thanks for any help you are able to give.
Best,
CMobley7
The text was updated successfully, but these errors were encountered: