[ros-users] move_arm, motion_planner, mechanism_controllers, ... lost :S

Antons Rebguns anton at email.arizona.edu
Fri Mar 4 18:44:23 UTC 2011


Matei,

> Yes, it does look like you have all components in. Let us know what
> kind of problems you run into when integrating them in the
> PickupAction that the manipulation pipeline serves, we should be able
> to help and then hopefully fix any wrinkles in the upcoming
> manipulation releases for diamondback. Anything that you find
> counter-intuitive or unclear will help is write better documentation
> and clearer API's.

Ok, good, I will try to integrate all the pieces today. Will let you
know if something doesn't make sense.

> What hand are you using?

DIY :) It's a custom 7DOF arm built from Robotis Dynamixel motors.

Thanks,
Anton

> On Fri, Mar 4, 2011 at 10:10 AM, Ugo Cupcic <ugo at shadowrobot.com> wrote:
>>
>>
>> On Fri, Mar 4, 2011 at 5:51 PM, Antons Rebguns <anton at email.arizona.edu>
>> wrote:
>>>
>>> Ugo,
>>>
>>> > I'll be happy to! Where are you at regarding the grasping pipeline?
>>>
>>> Well, I cac plan trajectories and execute them, I can segment objects
>>> from the floor and I hooked up OpenRAVE grasp planning to figure out
>>> the grasp pose for the gripper. Looking at the object_manipulator
>>> package I think I have all the required components that need to put
>>> together. Is that the package you are using?
>>
>> Yes I'm using that package. You seem to have everything ready for
>> integration then!
>>
>> If you do some object recognition after your segmentation, you should also
>> probably have a look at the household_objects node.
>>>
>>> > Things are not yet completely clear in my mind regarding the grasping
>>> > pipeline, so I haven't written a tutorial. But it's definitely in my
>>> > todo
>>> > list.
>>> >
>>> > Regarding my previous questions, I was in fact missing the
>>> > opml_planning. I
>>> > added it and it is now closer to be working ;)
>>>
>>> Good, glad to hear that. There's one problem I had with OMPL, that
>>> took me a few hours to figure out. It was always failing to plan
>>> saying that my current pose is in collision, which it definitely
>>> wasn't. It turned out it thought that 2 links of the arm were in
>>> collision even though they are pretty far away in that configuration.
>>> I had to disable collision checks between these two links. If anybody
>>> has an idea what might be happening there I'd like to hear it. Maybe
>>> there's a way to visualize the collision meshes that OMPL uses to
>>> figure out if current state is in collision.
>>>
>>> Anyway, you might not see that at all :)
>>>
>>> Thanks,
>>> Anton
>>>
>>> > On Fri, Mar 4, 2011 at 5:07 PM, Antons Rebguns <anton at email.arizona.edu>
>>> > wrote:
>>> >>
>>> >> Ugo,
>>> >>
>>> >> Let me know if you have any questions or need clarification on
>>> >> anything. I, on the other hand, trying to get the grasping pipeline
>>> >> working and I saw that you made some considerable progress there. Care
>>> >> to share you experiences?
>>> >>
>>> >> Thanks,
>>> >> Anton
>>> >>
>>> >> On Fri, Mar 4, 2011 at 9:53 AM, Ugo Cupcic <ugo at shadowrobot.com> wrote:
>>> >> > Hi,
>>> >> >
>>> >> > Yes I saw it a while ago, and had forgotten about this link. Thanks.
>>> >> >
>>> >> > I'll check your code and see if it helps me with my problems :)
>>> >> >
>>> >> > Cheers,
>>> >> >
>>> >> > Ugo
>>> >> >
>>> >> > On Fri, Mar 4, 2011 at 4:44 PM, Antons Rebguns
>>> >> > <anton at email.arizona.edu>
>>> >> > wrote:
>>> >> >>
>>> >> >> Ugo,
>>> >> >>
>>> >> >> Have you looked at this tutorial:
>>> >> >>
>>> >> >>
>>> >> >>
>>> >> >>
>>> >> >> http://www.ros.org/wiki/arm_navigation/Tutorials/Running%20arm%20navigation%20on%20non-PR2%20arm
>>> >> >>
>>> >> >> I outlined the steps I took to make my arm work with ROS motion
>>> >> >> planning pipeline.
>>> >> >>
>>> >> >> Anton
>>> >> >>
>>> >> >> On Fri, Mar 4, 2011 at 9:27 AM, Ugo Cupcic <ugo at shadowrobot.com>
>>> >> >> wrote:
>>> >> >> > Hi,
>>> >> >> >
>>> >> >> > I'm getting more and more lost trying to implement the move_arm
>>> >> >> > properly, so
>>> >> >> > sorry in advance if my questions are a bit messy.
>>> >> >> >
>>> >> >> > At first I duplicated the move_arm package and completely modified
>>> >> >> > it
>>> >> >> > for my
>>> >> >> > arm. It kind of worked but it didn't seem like a good solution, as
>>> >> >> > move_arm
>>> >> >> > seems to be very generic:
>>> >> >> > The only command output from this package seem to be a
>>> >> >> > pr2_controllers_msgs::JointTrajectoryAction. So if I implement my
>>> >> >> > own
>>> >> >> > joint_trajectory_action_controller.cpp, which is a
>>> >> >> > SimpleActionServer<pr2_controllers_msgs::JointTrajectoryAction>,
>>> >> >> > to
>>> >> >> > move
>>> >> >> > my
>>> >> >> > arm, will it work?
>>> >> >> >
>>> >> >> > To use the generic move_arm package, I'm still missing a
>>> >> >> > motion_planner
>>> >> >> > (I
>>> >> >> > think). Should I try to use the ompl_planner or the
>>> >> >> > ik_constrained_planner
>>> >> >> > or something else? When I'm trying to compile the
>>> >> >> > ik_constrained_planner
>>> >> >> > it's failing to build due to the generic_kinematics overlay (from
>>> >> >> >
>>> >> >> >
>>> >> >> >
>>> >> >> > https://code.ros.org/svn/wg-ros-pkg/stacks/kinematics/branches/generic_kinematics).
>>> >> >> >
>>> >> >> > Hope you can give me some hints on what I am doing wrong.
>>> >> >> >
>>> >> >> > Cheers,
>>> >> >> >
>>> >> >> > Ugo
>>> >> >> >
>>> >> >> > --
>>> >> >> > Ugo Cupcic | Shadow Robot Company | ugo at shadowrobot.com
>>> >> >> > Software Engineer | 251 Liverpool Road |
>>> >> >> > need a Hand? | London N1 1LX | +44 20 7700 2487
>>> >> >> > http://www.shadowrobot.com/hand/ @shadowrobot
>>> >> >> >
>>> >> >> >
>>> >> >> > _______________________________________________
>>> >> >> > ros-users mailing list
>>> >> >> > ros-users at code.ros.org
>>> >> >> > https://code.ros.org/mailman/listinfo/ros-users
>>> >> >> >
>>> >> >> >
>>> >> >
>>> >> >
>>> >> >
>>> >> > --
>>> >> > Ugo Cupcic | Shadow Robot Company | ugo at shadowrobot.com
>>> >> > Software Engineer | 251 Liverpool Road |
>>> >> > need a Hand? | London N1 1LX | +44 20 7700 2487
>>> >> > http://www.shadowrobot.com/hand/ @shadowrobot
>>> >> >
>>> >> >
>>> >> > _______________________________________________
>>> >> > ros-users mailing list
>>> >> > ros-users at code.ros.org
>>> >> > https://code.ros.org/mailman/listinfo/ros-users
>>> >> >
>>> >> >
>>> >
>>> >
>>> >
>>> > --
>>> > Ugo Cupcic | Shadow Robot Company | ugo at shadowrobot.com
>>> > Software Engineer | 251 Liverpool Road |
>>> > need a Hand? | London N1 1LX | +44 20 7700 2487
>>> > http://www.shadowrobot.com/hand/ @shadowrobot
>>> >
>>> >
>>> > _______________________________________________
>>> > ros-users mailing list
>>> > ros-users at code.ros.org
>>> > https://code.ros.org/mailman/listinfo/ros-users
>>> >
>>> >
>>
>>
>>
>> --
>> Ugo Cupcic | Shadow Robot Company | ugo at shadowrobot.com
>> Software Engineer | 251 Liverpool Road |
>> need a Hand? | London N1 1LX | +44 20 7700 2487
>> http://www.shadowrobot.com/hand/ @shadowrobot
>>
>>
>> _______________________________________________
>> ros-users mailing list
>> ros-users at code.ros.org
>> https://code.ros.org/mailman/listinfo/ros-users
>>
>>
>
>
>
> --
> Matei Ciocarlie
> Research Scientist
> Willow Garage Inc.
> 650-475-9780
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>



More information about the ros-users mailing list