Replies: 1 comment 7 replies
-
this is not a MoveIt issue. As we've discussed in Moveit & Singularities on ROS Answers, MoveIt produces As to the question per se: afaik, there is no direct, built-in support in MoveIt to make it avoid singularities. That's partly because it plans in joint space, and motions in joint space don't run into this problem. You could potentially implement a cost function (but that's an ongoing development) and/or implement a custom sampler (which would inspect the jacobian and reject samples which are close to a singularity to steer the planner(s) away from them). But both of those options require custom development. I believe this is actually an issue with the PFL / FSU configuration of the YRC1000u. Perhaps @ted-miller and/or @EricMarcil could clarify something here. |
Beta Was this translation helpful? Give feedback.
-
|
I have pinged a colleague who is the expert on HC robots to see if there is a way to suppress that alarm without disabling the PFL function. As gavanderhoorn said, singularities shouldn't be an issue when moving in joint-space. |
Beta Was this translation helpful? Give feedback.
-
|
It seems that the problem is not caused by the motion necessarily. From the HC documentation:
Please see section 4.3 (page 89) of 181437-1CD for the details of the conditions (1) & (2) along with images. |
Beta Was this translation helpful? Give feedback.
-
|
Is there a way to avoid conditions (1) and (2) while operating via MoveIt? |
Beta Was this translation helpful? Give feedback.
-
|
you could see whether a custom constraint sampler could help (unfortunately that page is a stub, there doesn't appear to be any more detailed information. Searching Github (like this) for existing implementations might be an alternative). It would 'inform' MoveIt about additional constraints it should take into account while planning motions. An alternative could be using MoveIt's support for cost-functions (which is under development right now in MoveIt 2), constraint manifolds (Planning with Approximated Constraint Manifolds for some documentation) and/or use OMPL constrained planning (not sure what the status of this is). Using OMPL Constrained Planning is an example of the latter (note: MoveIt 2). All of that would require custom development. I would suggest to post on the moveit/issues tracker, or the moveit2/issues tracker to discuss these options, as we don't have the expertise here to guide you. (if you do post there, please keep things connected and post a comment here with a link to your new issue(s)) Note that all of these options would probably result in you not being able to plan motions from and/or to the pose(s) you show(ed). Edit: you might be able to use positions and/or orientation constraints, but I'm not sure how well that would work. That would not require custom development (as in: a new plugin), as those are part of a MoveIt planning request. RViz however does not support them out-of-the-box. I'd still suggest asking on the MoveIt issue tracker. |
Beta Was this translation helpful? Give feedback.
-
|
I've opened an issue in the MoveIt issue tracker as suggested by @gavanderhoorn |
Beta Was this translation helpful? Give feedback.
-
Hi all,
I'm using #541 and MoveIt for controlling my HC10DTP motoman cobot.
After playing with this setup for a while I noticed that I'm running into singularities from time to time, meaning that my robot halts when it reaches a singularity and then I have to move it manually (via teach mode) out of that position and re-plan (sometimes multiple times).
I want to know if there is a way to tell MoveIt to avoid singularities.
The output from the console
======================================================
I'm adding some photos and a link to a video that shows the robot reaching a singularity point and halting:
YouTube video of the robot halting when it reaches a singularity
Current and goal state in Rviz:
Screenshots of the pendant after reaching a singularity:
Beta Was this translation helpful? Give feedback.
All reactions