Closed dudasdavid closed 3 years ago
For reference, my calibration_devel-staging
currently is the same as in #538.
Just to clarify: this repository is not maintained by Universal Robots. It's a community effort at this point.
As to the problem you describe: at this point I'm inclined to assume it's a problem in MoveIt.
MoveIt development is quite active, and the differences between Melodic and Noetic are significant (Noetic is essentially tracking master
, while Melodic is not).
From my own experience, "unable to sample any valid states for goal tree" is often caused by either:
My suggestions would be:
ur_kinematics
and whatever changed in Noetic in MoveIt, and see whether anything changesFor now I'm going to close this here, but if it turns out something needs to be fixed here, we'll re-open.
It was tested with KDL (since the UR IK doesn't really work well with limited joints and MoveIt). I can partially accept that it's a MoveIt issue, however the same MoveIt Commander application works with Panda robot in the simulation.
I'd also exclude a trivial MoveIt config issue, because I observe the same problem with @fmauch's calibration_devel
and calibration_devel-staging
branches.
@gavanderhoorn is there any planned official Universal Robots MoveIt support, if this repository isn't an official one anymore?
@gavanderhoorn is there any planned official Universal Robots MoveIt support, if this repository isn't an official one anymore?
It's not non-official, it is just not developed / maintained by the Company Universal Robots but driven by community effort.
is there any planned official Universal Robots MoveIt support, if this repository isn't an official one anymore?
this may have been unclear, but it's never been "official". It's always been a community effort. It's existed for far longer than UR's involvement with their driver, or with their interest in ROS actually.
And as I'm not part of UR, I'm not privy to what they plan to do or not.
As to your problem: this will require reproduction and then debugging MoveIt. That's unfortunately not something I can make time available for right now.
As a first step, I'd turn up the logging config for MoveIt itself to DEBUG
, which might provide more insight into where exactly something is failing.
Hi Universal Robots team,
I know that there isn't any official support yet for Noetic in this repository, but I can see that you're working on it in the Universal_Robots_ROS_Driver. @fmauch I'm using a fork based on your calibration_devel branch, but I tested your calibration_devel-staging, too.
I have no issues with ROS Melodic, only with Noetic and only when using moveit commander with pose target (using set_pose_target).
You can reproduce my issue with the following steps: 1)
git clone -b calibration_devel-staging https://github.com/fmauch/universal_robot
2) Changearm_controller
topos_joint_traj_controller
in ur3e_moveit_planning_execution.launch 3) Launch the following launchfiles:roslaunch ur_gazebo ur3e_bringup.launch limited:=true
roslaunch ur3e_moveit_config ur3e_moveit_planning_execution.launch sim:=true limited:=true
4) Run my example code that can be found here.This is the error I get from moveit:
The same moveit commander script works in ROS Melodic (with Python 2.7) and I'm using a similar one for weeks both in simulation and a real UR3e. By the way I have the exactly same issue with the real robot using Noetic, so it's not a Gazebo related problem.
Do you have any idea what can be wrong with MoveIt on Noetic?
Let me know if you have any question, I'm happy to test and/or give further information!
David