Store picking position as use for placement

Q&A
1

Dear MM-Team,
how could I achieve the following behaviour?

  1. Detect Grasp-Point via Visual Move (Call Vision A)
  2. Pick the item
  3. Do some robot movements with the item attached, call other vision project (Vision B, other camera)
  4. Drive back to the original picking position and detach the item
    (perfect placement, since it was the picking position in the first place)

How can I store the pickup position from the vision move and access it later on?
Ideally with a relative move beforehand, i.e. z=-0.3m.

I tried with the flags:

  • Reuse Vision Result
  • Share Vision Result
  • Keep Shared Vision Result When Succeed

and then just again a visual move (with the same vision project as the first step) for the placement,
but this didnt work and lead to strange behaviour.

Thank you for the help,

best,
Jonathan

2

The above requirements can be achieved through adapter.

3

Hi Jonathan,
You can achieve this by using adapter and External Move Step (Outer Move in ealier versions).
After the Visual Move (grasping point) planned, get the planned result using adapter, and set the waypoint to the Outer Move Step. After that, when executing the Outer Move Step, it can do the perfect placement.

For the External Move Step, You can refer to this document:
External Move (mech-mind.net)

Best,
Xinv

4

Hi Xinv,
thank you very much for your response.

In case someone needs this as well:

  • read poses_planned via VisionResultSelectedAtService
  • flip z-axis of quaternion
  • write outer_move via OuterMoveService as move_target_type = TCP_POSE
1 Like