You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, I am trying to get accurate move for ur5 MoveIt control based on the fact that each UR is calibrated individually and thus URDF is not accurate. This leads to the max error of 3 mm.
For now, I have used tool0_controller to read real end-effector pose successfully. However, it was not enough to just read it for getting accurate move target position. Is there any way to control that frame instead of ee-link or tool0 via MoveIt?
From #16 , I also extracted the calibrated DH parameters from a UR robot. But I am not quite understand the meaning of some parameters, like joint_checksum, joint_raw_offset, joint_selftest_data_crc. Which parameter should I use to get an accurate ur5.urdf? @sepjansen
If you understand DH parameters, the values you show should allow you to update your urdf and get as close as possible to 'true' values with just the urdf. Beyond that, tool0_controller is available.
For now, I have used tool0_controller to read real end-effector pose successfully. However, it was not enough to just read it for getting accurate move target position. Is there any way to control that frame instead of ee-link or tool0 via MoveIt?
I'm not sure I understand you here: the driver controls the robot in joint space, so at first glance, the answer would be: "no, you cannot control tool0_controller". However, it should be possible to write a Cartesian controller for ros_control. That controller could then close the loop over any arbitrary TF frame, including tool0_controller, and use either the position or velocity interfaces to implement a position control loop that applies the necessary corrections.
While this is possible, beware of some of the current issues, such as #139, #132 and #114.
Hi, I am trying to get accurate move for ur5 MoveIt control based on the fact that each UR is calibrated individually and thus URDF is not accurate. This leads to the max error of 3 mm.
For now, I have used
tool0_controller
to read real end-effector pose successfully. However, it was not enough to just read it for getting accurate move target position. Is there any way to control that frame instead ofee-link
ortool0
via MoveIt?From #16 , I also extracted the calibrated DH parameters from a UR robot. But I am not quite understand the meaning of some parameters, like
joint_checksum
,joint_raw_offset
,joint_selftest_data_crc
. Which parameter should I use to get an accurate ur5.urdf? @sepjansenCalibration.conf file:
Thanks!
The text was updated successfully, but these errors were encountered: