Using ABB robotstudio's Virtual Controller for testing with ROS

[Simon Jansen]

Hey all,

We're busy building a demonstration of ROS industrial on an ABB IRB6640 robot. Since we do not have a lot of access to the actual hardware, we tried to test the communication in ABB RobotStudio and it worked suprisingly well. One PC is running windows with a virtual model of the robot in RobotStudio, the other PC is running ubuntu with the ROS industrial interface, moveit, etc.

It gave us valuable testing time and the possibility to test with all types of ABB robot (not just the IRB6640) and we can play around with different grippers and other I/O interfaces.

I did not see any mention of this on the wiki or this mail group , so I'm just throwing the idea out there for anyone working with ROS industrial and an ABB robot.

We documented the steps to configure the controller in robotstudio and are glad to share it with anyone interested in this approach.

regards,

Simon Jansen

Alten Mechatronics

[Shaun Edwards]

Simon,

Thanks for information.  ABB deserves credit for creating a high quality simulation (all the way down to TCP socket calls).  I think this information would make sense on the wiki.  You can send me the info and I can add it to the wiki, or feel free to add it yourself.

Thanks again,

Shaun

--
You received this message because you are subscribed to the Google Groups "swri-ros-pkg-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to swri-ros-pkg-d...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Shaun Edwards]

Simon,

Thanks for information.  ABB deserves credit for creating a high quality simulation (all the way down to TCP socket calls).  I think this information would make sense on the wiki.  You can send me the info and I can add it to the wiki, or feel free to add it yourself.

Thanks again,

Shaun

On Wed, Sep 25, 2013 at 9:48 AM, Simon Jansen <seps...@gmail.com> wrote:

--

[jose brizuela]

Hello Simon, this aproach is very interesting, probably we can continue working on this to experience other funcionalities. Could you share the documentation of the configuration of robotstudio and the tests?.

Thanks in advance.

Regards.

[Simon Jansen]

Hello Jose,

The documentation can be found with the other ABB tutorials on the ROS wiki:

http://wiki.ros.org/abb/Tutorials/RobotStudio

Note that this only documents the configuration in robotstudio: testing should be the same as with a real robot.

Let me know if you encounter any problems with it.

regards,

Simon

--

[Kathrin]

Dear Simon,

i was configuring and running the abb node in robot studio, which is working fine, thank you for the tutorial.

I have one question concerning the abb ros node: why would the abb hardware configuration not allow position or trajectory streaming, but only trajectory downloading? Or is this only not yet implemented, and the hardware or the controller would allow this?

Thank you very much in advance for your answer,

Regards,

Kathrin

[Simon Jansen]

Hello Kathrin,

I did not actually write the code running on the ABB controller and I don't know why the choice was made for trajectory downloading rather than streaming. You can have a look at the RAPID code yourself, if you're familiar with programming ABB robots directly:

https://github.com/ros-industrial/abb/tree/groovy-devel/abb_common/rapid

I think there is some general discussion about position streaming within ROS-I, perhaps Shaun Edwards can be more helpfull here.

regards,

Simon

--

[jz...@swri.org]

Kathrin,

 I wrote the ABB driver code, based heavily on original work done by Ed Venator at Case Western.  The download-vs-streaming decision was made primarily because that approach is easier and more stable for most robots.

 In some of our original ROS-I discussions, we recognized that the decision for whether to implement download or streaming interfaces may be dictated by robot-language limitations or application needs (or both).  The intent was that we might eventually provide both methods for robot platforms, and let the user choose based on their application needs.  However, so far, we have mostly implemented what was needed for our applications and have not taken the time to develop alternative position-control methods.  We certainly welcome any contributions, if you’d like to expand / adapt the ABB driver to support position-streaming control methods!

 Industrial robots are typically designed to execute an entire trajectory of known positions, rather than processing incoming position commands on-the-fly.  This is why we often implement a “trajectory download” approach first, because it is a close match to the robot’s native motion processing.  We have successfully implemented motion-streaming interfaces on an industrial robot before (see the original motoman dx100 driver), but this method often requires “hacks”, unless there is support for this functionality in the robot’s native motion-control language.

 For an ABB, it might be possible to use an RSI hook to implement the streaming position commands, but trying to execute streaming commands using standard “MoveAbsJ” commands may result in poor motion performance.  In particular, the robot’s motion processor will be constantly looking ahead to the upcoming motion commands to determine whether the robot should decelerate to a stop or blend the current trajectory into the next motion command.  You would need to make sure that there is always a sufficient “buffer” of upcoming position commands to ensure that the motion processor does not try to decelerate the robot in the middle of a streamed trajectory.  The details of this look-ahead buffer are often not well-documented, and must be determined through trial and error.

 Hope this helps!
  Jeremy Zoss
  Southwest Research Institute

-----------------

[Kathrin]

Hi Jeremy,

thank you very much for your quick reply, just for your information, I am very new to ROS, so my questions might be very basic.

Basically, what I already did in a previous work, (without ROS), was streaming positions (frames, not joint positions) to the ABB controller over a socket, writing them into a buffer of 3-5, and reading the positions out of this buffer, in order for the robot to be able to interpolate in between the different frames, as you described above. That is why I am wondering that the ABB ROS Node in its current state only implemented the joint trajectory downloading, while other robots would support the position streaming, like the UR's (is that right?) or motoman robots. If this is just a matter of implementation I might have a closer look into this.

My other question concerns the joint positions in general: was it never planned that trajectories can be transmitted either over joint positions or within the Cartesian coordinate system with frames? Does this have a special reason, or do you actually don't need this for your applications? (I am mainly asking this because something came into my mind: as far as I know, Kuka robot controllers, for instance, (I don't know enough about Abb's) are correcting their trajectories according to their deflection, since their joints are quite soft, and therefore heavy load at the end-effector has to be taken into consideration. This is not taken into account when executing given joint positions, only when executing given trajectories with frames. In case I am solving the inverse kinematics outside of the controller it is not possible to implement this position-correction within the kinematic model, which, in the end, will cause a certain imprecision in the execution of trajectories?)

And I found no information concerning the ABB RSI hook that you mentioned, could you maybe explain this also?

Thank you very much in advance,

Regards,

Kathrin

[Kathrin]

Hi Simon,

thank you for your reply! I already got an answer below from Jeremy, so that already clarified some things.

I am trying to use ROS for the online control of an ABB IRB4600. Could you tell me which ROS nodes you are using for the path planning, are you using the moveit package?

Thanks a lot,

Kathrin

[jz...@swri.org]

Kathrin,

The method you describe for buffering streaming position data on the ABB controller is the approach we have used on other robots.  I describe this method as a "hack", because it is very dependent on the robot's internal look-ahead motion processing.  Depending on the trajectory being executed (point spacing, motion speeds, etc), the resulting motion performance (speed, path "smoothness", TCP position near path-segment transitions, etc.) may not match what you expect.  In contrast, the trajectory-download method provides predictable results, although the streaming functionality is lost.

Some of the existing streaming-position robot drivers (e.g. motoman's MotoROS code) utilize robot-specific motion commands to bypass the normal look-ahead motion processing and issue the motion commands "directly" to the robot.  These interfaces were provided by the robot manufacturers to support on-the-fly motion commands, and do not suffer the same types of look-ahead performance issues described above.  We try to take advantage of these interfaces where available, as they typically result in better performance than a look-ahead buffer-based approach.

Unfortunately, I had a moment of distraction when writing the earlier reply.  The RSI interface I mentioned is available on Kuka robots, not ABB robots.  Sorry for the confusion!

Feel free to implement a similar streaming interface to what you have used before.  It's possible that this works smoothly on an ABB robot.  I have not tried this, myself, so I don't know how the robot handles this type of streaming motion commands.  Even if it is not an "ideal" implementation, it provides a capability not supported by ROS-I, and we would welcome your contribution!

As far as joint-vs-cartesian motion commands go, we have traditionally implemented joint-based command interfaces, because that is how ROS typically interfaces with robots.  Most (all?) of the current ROS planners work in joint-space, and generate joint trajectories for execution on the robots.  We recognize that this bypasses some of the functionality present on the robot controllers (IK).  However, trying to send cartesian commands brings its own set of challenges in how to transmit cartesian-related concepts (tool / user frames, arm "configuration" flags, etc.) in a platform-independent manner.  We are planning to eventually provide support for cartesian-based motion control (see the future section of the ROS-I spec), but this is not on the near-term roadmap.

[Kathrin Dörfler]

Dear Jeremy,

thank you very much for your support. I now understand the "hack" :) If I come to it, I will look into the streaming capabilities, and keep you up to date on any progress. I also see the point of "platform-independency" in the case of joint-vs-cartesian motion commands.

Thank you very much again,

Kathrin

2013/10/23 <jz...@swri.org>

--
You received this message because you are subscribed to a topic in the Google Groups "swri-ros-pkg-dev" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/swri-ros-pkg-dev/uDu7O-18Z1Q/unsubscribe.
To unsubscribe from this group and all its topics, send an email to swri-ros-pkg-d...@googlegroups.com.

[Simon Jansen]

Hi Kathrin,

We used indeed the moveit package for path planning.

There's a very nice wiki on how to create a model of your robot (there is currently not 1 made yet for the irb4600, so you'll have to do it yourself) and how to create a moveit package for it:

http://wiki.ros.org/Industrial/Tutorials/Create_a_MoveIt_Pkg_for_an_Industrial_Robot

You can also have a look at existing moveit packages, like the one for the ABB IRB2400:

https://github.com/ros-industrial/abb/tree/groovy-devel/irb_2400_moveit_config

Hope this helps you to get started.

regards,

Simon

[Kathrin Dörfler]

Dear Simon,

thank you very much for your reply! Ok, then I will try to get started with this, or do you know also about other options?

One more question: are you running the ros on groovy or hydro, and which one would you recommend?

Thanks a lot,

Kathrin

2013/10/24 Simon Jansen <seps...@gmail.com>

--
You received this message because you are subscribed to a topic in the Google Groups "swri-ros-pkg-dev" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/swri-ros-pkg-dev/uDu7O-18Z1Q/unsubscribe.
To unsubscribe from this group and all its topics, send an email to swri-ros-pkg-d...@googlegroups.com.

[Simon Jansen]

Hello Kathrin,

I think as far as path planning in ROS goes, MoveIt is pretty much the way to go

We recently switched to hydro, because moveit is still developing fast and the hydro version gives you the latest features.

However, there is no hydro release yet of the ROS Industrial stack, but it should be soon. Can anyone confirm that by the way?

regards,

Simon

[Shaun Edwards]

Yes. Hydro will be released soon. Probably before the end of the year.

Please submit issues/pull requests for any changes that you guys find. This will help speed along the process. 

-Shaun. 

Sent from my mobile phone

[Brian O'Neil]

Hi Kathrin,

The current GitHub source for ROS-I is for hydro and above. You can install from debs for groovy, or the source is still available on the old google code repo. MoveIt on groovy is much more stable, but the hydro release offers considerably more features. Also, there are enough changes between MoveIt in groovy and hydro that if you devote a lot of time to writing groovy code, there will be some significant retooling required when you move to hydro. I would recommend starting with hydro. Also, join the MoveIt mailing list and ask questions there. Ioan is a champ about responding to MoveIt related questions, and the community there seems to be growing fast.

--Brian

Brian O'Neil
Posdoctoral Research Associate
Nuclear Robotics Program

University of Texas at Austin

Phone: (480) 363-8963

[Kathrin]

Hi Brian,

I now switched to hydro, thank you for your information!

Regards,

Kathrin

[Kathrin]

Thank you Simon,

I am setting up the hydro release now. Thank you for your info!

Kathrin

[Kathrin]

Hi Simon,

I am stuck, maybe you know sthg: I set up hydro and tried to build the ros-Industrial stacks in a catkin-workspace, which is not working out...  (e.g. cannot get the abb_common launch file to run: Error: ERROR: cannot launch node of type [abb_common/robot_state]: can't locate node [robot_state] in package [abb_common]

ERROR: cannot launch node of type [abb_common/motion_download_interface]: can't locate node [motion_download_interface] in package [abb_common])

What I did: I set up a catkin_ws into which src folder I cloned the github repositories from ROS-I. Then I did catkin_make_isolated in the catkin_ws, which was executed with no errors. Still I cannot run the launch files or rosnodes for some reason... You might know what I did wrong?

Thanks so much again,

Kathrin

[G.A. vd. Hoorn - 3ME]

On 29/10/2013 14:31, Kathrin wrote:
> Hi Simon,
>
> I am stuck, maybe you know sthg: I set up hydro and tried to build the
> ros-Industrial stacks in a catkin-workspace, which is not working out...
> (e.g. cannot get the abb_common launch file to run: Error: ERROR: cannot
> launch node of type [abb_common/robot_state]: can't locate node
> [robot_state] in package [abb_common]
> ERROR: cannot launch node of type [abb_common/motion_download_interface]:
> can't locate node [motion_download_interface] in package [abb_common])
>
> What I did: I set up a catkin_ws into which src folder I cloned the github
> repositories from ROS-I. Then I did catkin_make_isolated in the catkin_ws,
> which was executed with no errors. Still I cannot run the launch files or
> rosnodes for some reason... You might know what I did wrong?

(I'm not Simon, but I might be able to help)

did you source the setup.bash in the 'catkin_ws/devel' dir (I tend to do
this in a new terminal)? Otherwise roslaunch/rosrun will only look in
the src dir.


Gijs

> Am Donnerstag, 24. Oktober 2013 10:46:30 UTC+2 schrieb Simon Jansen:
>>
>> Hello Kathrin,
>>
>> I think as far as path planning in ROS goes, MoveIt is pretty much the way
>> to go
>>
>> We recently switched to hydro, because moveit is still developing fast and
>> the hydro version gives you the latest features.
>>
>> However, there is no hydro release yet of the ROS Industrial stack, but it
>> should be soon. Can anyone confirm that by the way?
>>
>> regards,
>> Simon
>>
>>

>> On Thu, Oct 24, 2013 at 10:37 AM, Kathrin D�rfler<doerfler...@gmail.com<javascript:>

>>> wrote:
>>
>>> Dear Simon,
>>>
>>> thank you very much for your reply! Ok, then I will try to get started
>>> with this, or do you know also about other options?
>>>
>>> One more question: are you running the ros on groovy or hydro, and which
>>> one would you recommend?
>>>
>>> Thanks a lot,
>>> Kathrin
>>>
>>>

>>> 2013/10/24 Simon Jansen<seps...@gmail.com<javascript:>>

>>>
>>>> Hi Kathrin,
>>>>
>>>> We used indeed the moveit package for path planning.
>>>> There's a very nice wiki on how to create a model of your robot (there
>>>> is currently not 1 made yet for the irb4600, so you'll have to do it
>>>> yourself) and how to create a moveit package for it:
>>>>
>>>> http://wiki.ros.org/Industrial/Tutorials/Create_a_MoveIt_Pkg_for_an_Industrial_Robot
>>>>
>>>> You can also have a look at existing moveit packages, like the one for
>>>> the ABB IRB2400:
>>>>
>>>> https://github.com/ros-industrial/abb/tree/groovy-devel/irb_2400_moveit_config
>>>>
>>>> Hope this helps you to get started.
>>>>
>>>> regards,
>>>> Simon
>>>>
>>>>

>>>> On Wed, Oct 23, 2013 at 5:50 PM, Kathrin<doerfler...@gmail.com<javascript:>

>>>>> wrote:
>>>>
>>>>> Hi Simon,
>>>>>
>>>>> thank you for your reply! I already got an answer below from Jeremy, so
>>>>> that already clarified some things.
>>>>> I am trying to use ROS for the online control of an ABB IRB4600. Could
>>>>> you tell me which ROS nodes you are using for the path planning, are you
>>>>> using the moveit package?
>>>>>
>>>>> Thanks a lot,
>>>>> Kathrin
>>>>>
>>>>>
>>>>>
>>>>> Am Mittwoch, 23. Oktober 2013 09:16:35 UTC+2 schrieb Simon Jansen:
>>>>>>
>>>>>> Hello Kathrin,
>>>>>>
>>>>>> I did not actually write the code running on the ABB controller and I
>>>>>> don't know why the choice was made for trajectory downloading rather than
>>>>>> streaming. You can have a look at the RAPID code yourself, if you're
>>>>>> familiar with programming ABB robots directly:

>>>>>> https://github.com/ros-**industrial/abb/tree/groovy-**
>>>>>> devel/abb_common/rapid<https://github.com/ros-industrial/abb/tree/groovy-devel/abb_common/rapid>

>>>>>>> send an email to swri-ros-pkg-d...@**googlegroups.com.
>>>>>>>
>>>>>>> For more options, visit https://groups.google.com/**groups/opt_out<https://groups.google.com/groups/opt_out>
>>>>>>> .

>>>>>>>
>>>>>>
>>>>>> --
>>>>> You received this message because you are subscribed to the Google
>>>>> Groups "swri-ros-pkg-dev" group.
>>>>> To unsubscribe from this group and stop receiving emails from it, send

>>>>> an email to swri-ros-pkg-d...@googlegroups.com<javascript:>.

>>>>>
>>>>> For more options, visit https://groups.google.com/groups/opt_out.
>>>>>
>>>>
>>>> --
>>>> You received this message because you are subscribed to a topic in the
>>>> Google Groups "swri-ros-pkg-dev" group.
>>>> To unsubscribe from this topic, visit
>>>> https://groups.google.com/d/topic/swri-ros-pkg-dev/uDu7O-18Z1Q/unsubscribe
>>>> .
>>>> To unsubscribe from this group and all its topics, send an email to

>>>> swri-ros-pkg-d...@googlegroups.com<javascript:>.

>>>>
>>>> For more options, visit https://groups.google.com/groups/opt_out.
>>>>
>>>
>>> --
>>> You received this message because you are subscribed to the Google Groups
>>> "swri-ros-pkg-dev" group.
>>> To unsubscribe from this group and stop receiving emails from it, send an

>>> email to swri-ros-pkg-d...@googlegroups.com<javascript:>.

[Simon Jansen]

In addition to the answer of Gijs:

If this is the first time you setup a catkin workspace, you may find some hints on http://answers.ros.org from other ROS users (look for catkin or similar). I installed and compiled the abb package in hydro just fine (apart from some warnings about deprecated functions), so it seems to me a general catkin issue.

On Thu, Oct 24, 2013 at 10:37 AM, Kathrin Dörfler<doerfler...@gmail.com<javascript:>

To unsubscribe from this group and stop receiving emails from it, send an email to swri-ros-pkg-dev+unsubscribe@googlegroups.com.

[Kathrin]

Am Dienstag, 29. Oktober 2013 14:39:41 UTC+1 schrieb gavanderhoorn:

On 29/10/2013 14:31, Kathrin wrote:
> Hi Simon,
>
> I am stuck, maybe you know sthg: I set up hydro and tried to build the
> ros-Industrial stacks in a catkin-workspace, which is not working out...
>   (e.g. cannot get the abb_common launch file to run: Error: ERROR: cannot
> launch node of type [abb_common/robot_state]: can't locate node
> [robot_state] in package [abb_common]
> ERROR: cannot launch node of type [abb_common/motion_download_interface]:
> can't locate node [motion_download_interface] in package [abb_common])
>
> What I did: I set up a catkin_ws into which src folder I cloned the github
> repositories from ROS-I. Then I did catkin_make_isolated in the catkin_ws,
> which was executed with no errors. Still I cannot run the launch files or
> rosnodes for some reason... You might know what I did wrong?

(I'm not Simon, but I might be able to help)

did you source the setup.bash in the 'catkin_ws/devel' dir (I tend to do
this in a new terminal)? Otherwise roslaunch/rosrun will only look in
the src dir.


Gijs

Hi Gijs,

thank you for your reply. I was redoing everything, the second time it was working, I actually don`t know anymore, what the mistake was, it might really be that I forgot to source the setup.bash file.

I still got one Error in the dependencies: ERROR: the following packages/stacks could not have their rosdep keys resolved

to system dependencies:

irb_2400_arm_navigation: Cannot locate rosdep definition for [move_arm_warehouse]

irb_5400_arm_navigation: Cannot locate rosdep definition for [move_arm_warehouse]

but I could run catkin_make without further errors. Maybe sb knows an answer to this.

I have one further question (sorry if this might be stupid again), but am I supposed to locate the directories of this ABB Ros-Industrial Stack within the "catkin_ws/devel" directory, because I have them in my home folder now, where I created a catkin_ws for this only?

Thank you for your help

Kathrin

 

> Am Donnerstag, 24. Oktober 2013 10:46:30 UTC+2 schrieb Simon Jansen:
>>
>> Hello Kathrin,
>>
>> I think as far as path planning in ROS goes, MoveIt is pretty much the way
>> to go
>>
>> We recently switched to hydro, because moveit is still developing fast and
>> the hydro version gives you the latest features.
>>
>> However, there is no hydro release yet of the ROS Industrial stack, but it
>> should be soon. Can anyone confirm that by the way?
>>
>> regards,
>> Simon
>>
>>

>> On Thu, Oct 24, 2013 at 10:37 AM, Kathrin D�rfler<doerfler...@gmail.com<javascript:>

[Kathrin]

Hi Simon,

thank you for this hint, I guess I should have read some answers before...

Thanks,

Kathrin

To unsubscribe from this group and stop receiving emails from it, send an email to swri-ros-pkg-d...@googlegroups.com.

[G.A. vd. Hoorn - 3ME]

This is most likely caused by the fact that arm_navigation is no longer
available in Hydro. Those two packages depend on it (note the
'_arm_navigation' suffix). Since you will be using the MoveIt packages,
this should not pose a problem.

> I have one further question (sorry if this might be stupid again), but am I
> supposed to locate the directories of this ABB Ros-Industrial Stack within
> the "catkin_ws/devel" directory, because I have them in my home folder now,
> where I created a catkin_ws for this only?

No, the 'devel' directory gets populated with the necessary files during
building. You put all your sources in the src dir, then do a
'catkin_make' in your workspace dir (one level above the src dir), and
the devel directory contains the *result* of 'catkin_make' [*]. ROS
takes care of putting the right files in the right places. Just source
the devel-space setup.bash and everything should work (in that terminal).

> Thank you for your help
> Kathrin

Gijs

[*] Actually, the 'build' dir contains the build artefacts (binaries,
libraries, etc), which get either copied or symlinked into the devel dir
(or 'space', to follow proper ROS-nomenclature).

>>> Am Donnerstag, 24. Oktober 2013 10:46:30 UTC+2 schrieb Simon Jansen:
>>>> Hello Kathrin,
>>>>
>>>> I think as far as path planning in ROS goes, MoveIt is pretty much the
>> way
>>>> to go
>>>>
>>>> We recently switched to hydro, because moveit is still developing fast
>> and
>>>> the hydro version gives you the latest features.
>>>>
>>>> However, there is no hydro release yet of the ROS Industrial stack, but
>> it
>>>> should be soon. Can anyone confirm that by the way?
>>>>
>>>> regards,
>>>> Simon
>>>>
>>>>

>>>> On Thu, Oct 24, 2013 at 10:37 AM, Kathrin D�rfler<

[Kathrin]

and do you have several catkin workspaces with different packages, or are you supposed to have just one, with all your packages in there?

thanks, Kathrin

 

> Thank you for your help
> Kathrin


Gijs

[*] Actually, the 'build' dir contains the build artefacts (binaries,
libraries, etc), which get either copied or symlinked into the devel dir
(or 'space', to follow proper ROS-nomenclature).


>>> Am Donnerstag, 24. Oktober 2013 10:46:30 UTC+2 schrieb Simon Jansen:
>>>> Hello Kathrin,
>>>>
>>>> I think as far as path planning in ROS goes, MoveIt is pretty much the
>> way
>>>> to go
>>>>
>>>> We recently switched to hydro, because moveit is still developing fast
>> and
>>>> the hydro version gives you the latest features.
>>>>
>>>> However, there is no hydro release yet of the ROS Industrial stack, but
>> it
>>>> should be soon. Can anyone confirm that by the way?
>>>>
>>>> regards,
>>>> Simon
>>>>
>>>>

>>>> On Thu, Oct 24, 2013 at 10:37 AM, Kathrin D�rfler<

[G.A. vd. Hoorn - 3ME]

This mostly comes down to personal preference: workspaces can be
'chained', with one workspace layed 'on top' of another (ad infinitum).
There are some constraints to this, and it changes your workflow, but in
general you can do what works for you.

But as Simon mentioned earlier, these questions seem a bit off-topic for
this list, as they are more general catkin/ros/build-system related than
ROS-I.

>> Gijs
>>
>> [*] Actually, the 'build' dir contains the build artefacts (binaries,
>> libraries, etc), which get either copied or symlinked into the devel dir
>> (or 'space', to follow proper ROS-nomenclature).
>>
>>
>>>>> Am Donnerstag, 24. Oktober 2013 10:46:30 UTC+2 schrieb Simon Jansen:
>>>>>> Hello Kathrin,
>>>>>>
>>>>>> I think as far as path planning in ROS goes, MoveIt is pretty much
>> the
>>>> way
>>>>>> to go
>>>>>>
>>>>>> We recently switched to hydro, because moveit is still developing
>> fast
>>>> and
>>>>>> the hydro version gives you the latest features.
>>>>>>
>>>>>> However, there is no hydro release yet of the ROS Industrial stack,
>> but
>>>> it
>>>>>> should be soon. Can anyone confirm that by the way?
>>>>>>
>>>>>> regards,
>>>>>> Simon
>>>>>>
>>>>>>

>>>>>> On Thu, Oct 24, 2013 at 10:37 AM, Kathrin D�rfler<

[Brian O'Neil]

I don't think there's really a right or wrong way to manage your catkin workspaces. It is much simpler to operate out of a single workspace, and the drawbacks are minor. For example, your code will take a little longer to build because catkin will check everything in the workspace every time you run catkin_make.

The only time you must have multiple workspaces is if you want to use different versions of the same package, and then you have to be careful about the order in which you source the setup.bash files in the various workspaces. For example, if you installed moveit from debians, and later decided you wanted to check out the most recent source install, you could put the moveit source in a separate workspace and only source that workspace when you want to use the source install. This is known as overlaying a workspace and is discussed in more detail here. Pay attention to section 5 as it describes some pitfalls of chaining multiple workspaces.

Brian O'Neil
Posdoctoral Research Associate
Nuclear Robotics Program

University of Texas at Austin

Phone: (480) 363-8963

To unsubscribe from this group and stop receiving emails from it, send an email to swri-ros-pkg-d...@googlegroups.com.

[Kathrin Dörfler]

Hi Gijs,

thanks! Sorry for the misplacement...

Greetings,

Kathrin

2013/10/31 G.A. vd. Hoorn - 3ME <g.a.van...@tudelft.nl>

To unsubscribe from this group and all its topics, send an email to swri-ros-pkg-dev+unsubscribe@googlegroups.com.

[G.A. vd. Hoorn - 3ME]

On 31/10/2013 16:55, Kathrin Dörfler wrote:
> Hi Gijs,
> thanks! Sorry for the misplacement...

No worries, my comment wasn't meant as a
you-can't-send-these-questions-to-this-list per se, I just felt there
are mailinglists (and wiki pages) that can help you more efficiently.

See the response from Brian O'Neil for instance.

> 2013/10/31 G.A. vd. Hoorn - 3ME<g.a.van...@tudelft.nl>
>
>> On 31/10/2013 14:54, Kathrin wrote:
>>
>>> Am Dienstag, 29. Oktober 2013 19:49:44 UTC+1 schrieb gavanderhoorn:
>>>
>>>>
>>>> Kathrin wrote:
>>>>
>>>>> Am Dienstag, 29. Oktober 2013 14:39:41 UTC+1 schrieb gavanderhoorn:
>>>>>
>>>>>> On 29/10/2013 14:31, Kathrin wrote:
>>>>>>
>>>>>>> Hi Simon,
>>>>>>>
>>>>>>> I am stuck, maybe you know sthg: I set up hydro and tried to build the
>>>>>>> ros-Industrial stacks in a catkin-workspace, which is not working
>>>>>>>
>>>>>> out...
>>>>
>>>>> (e.g. cannot get the abb_common launch file to run: Error: ERROR:
>>>>>>>
>>>>>> cannot
>>>>>>
>>>>>>> launch node of type [abb_common/robot_state]: can't locate node
>>>>>>> [robot_state] in package [abb_common]
>>>>>>> ERROR: cannot launch node of type
>>>>>>>

>>>>>> [abb_common/motion_download_**interface]:

>>>>>>> doerfler...@gmail.com<**javascript:>

>>>>>>
>>>>>>> wrote:
>>>>>>>>> Dear Simon,
>>>>>>>>>
>>>>>>>>> thank you very much for your reply! Ok, then I will try to get
>>>>>>>>>
>>>>>>>> started
>>>>
>>>>> with this, or do you know also about other options?
>>>>>>>>>
>>>>>>>>> One more question: are you running the ros on groovy or hydro, and
>>>>>>>>>
>>>>>>>> which
>>>>>>
>>>>>>> one would you recommend?
>>>>>>>>>
>>>>>>>>> Thanks a lot,
>>>>>>>>> Kathrin
>>>>>>>>>
>>>>>>>>>

>>>>>>>>> 2013/10/24 Simon Jansen<seps...@gmail.com<**javascript:>>

>>>>>>>>>
>>>>>>>>> Hi Kathrin,
>>>>>>>>>>
>>>>>>>>>> We used indeed the moveit package for path planning.
>>>>>>>>>> There's a very nice wiki on how to create a model of your robot
>>>>>>>>>>
>>>>>>>>> (there
>>>>>>
>>>>>>> is currently not 1 made yet for the irb4600, so you'll have to do
>>>>>>>>>>
>>>>>>>>> it
>>>>
>>>>> yourself) and how to create a moveit package for it:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>

>>>>>> http://wiki.ros.org/**Industrial/Tutorials/Create_a_**
>>>> MoveIt_Pkg_for_an_Industrial_**Robot<http://wiki.ros.org/Industrial/Tutorials/Create_a_MoveIt_Pkg_for_an_Industrial_Robot>

>>>>
>>>>> You can also have a look at existing moveit packages, like the one
>>>>>>>>>>
>>>>>>>>> for
>>>>>>
>>>>>>> the ABB IRB2400:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>

>>>>>> https://github.com/ros-**industrial/abb/tree/groovy-**
>>>> devel/irb_2400_moveit_config<https://github.com/ros-industrial/abb/tree/groovy-devel/irb_2400_moveit_config>

>>>>
>>>>> Hope this helps you to get started.
>>>>>>>>>>
>>>>>>>>>> regards,
>>>>>>>>>> Simon
>>>>>>>>>>
>>>>>>>>>>

>>>>>>>>>> On Wed, Oct 23, 2013 at 5:50 PM, Kathrin<doerfler...@gmail.com<**

>>>>>>>>>>>> https://github.com/ros-****industrial/abb/tree/groovy-**<https://github.com/ros-**industrial/abb/tree/groovy-**>
>>>>>>>>>>>> devel/abb_common/rapid<

>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>> https://github.com/ros-**industrial/abb/tree/groovy-**
>>>> devel/abb_common/rapid<https://github.com/ros-industrial/abb/tree/groovy-devel/abb_common/rapid>
>>>>>
>>>>
>>>>> I think there is some general discussion about position streaming
>>>>>>>>>>>> within ROS-I, perhaps Shaun Edwards can be more helpfull here.
>>>>>>>>>>>>
>>>>>>>>>>>> regards,
>>>>>>>>>>>> Simon
>>>>>>>>>>>>
>>>>>>>>>>>>

>>>>>>>>>>>> On Tue, Oct 22, 2013 at 6:18 PM, Kathrin<doerfler...@gmail.com>*
>>>>>>>>>>>> *wrote:

>>>>> send an email to swri-ros-pkg-d...@**googlegrou**ps.com<http://googlegroups.com>

>>>>>>>>>>>>> .
>>>>>>>>>>>>>
>>>>>>>>>>>>> For more options, visit
>>>>>>>>>>>>>

>>>>>>>>>>>> https://groups.google.com/****groups/opt_out<https://groups.google.com/**groups/opt_out>

>>>>
>>>>> <https://groups.google.com/**groups/opt_out<https://groups.google.com/groups/opt_out>
>>>>>>>
>>>>>>
>>>>>>> .
>>>>>>>>>>>>>
>>>>>>>>>>>>> --
>>>>>>>>>>>>
>>>>>>>>>>> You received this message because you are subscribed to the Google
>>>>>>>>>>> Groups "swri-ros-pkg-dev" group.
>>>>>>>>>>> To unsubscribe from this group and stop receiving emails from it,
>>>>>>>>>>>
>>>>>>>>>> send
>>>>>>

>>>>>>> an email to swri-ros-pkg-d...@**googlegroups.com<swri-ros-pkg-d...@googlegroups.com>
>>>>>>>>>>> <javascript:>.
>>>>>>>>>>>
>>>>>>>>>>> For more options, visit https://groups.google.com/**
>>>>>>>>>>> groups/opt_out<https://groups.google.com/groups/opt_out>.
>>>>>>>>>>>
>>>>>>>>>>> --

>>>>>>>>>> You received this message because you are subscribed to a topic in
>>>>>>>>>>
>>>>>>>>> the
>>>>>>
>>>>>>> Google Groups "swri-ros-pkg-dev" group.
>>>>>>>>>> To unsubscribe from this topic, visit
>>>>>>>>>>
>>>>>>>>>>

>>>>>> https://groups.google.com/d/**topic/swri-ros-pkg-dev/uDu7O-**
>>>> 18Z1Q/unsubscribe<https://groups.google.com/d/topic/swri-ros-pkg-dev/uDu7O-18Z1Q/unsubscribe>

>>>>
>>>>> .
>>>>>>>>>> To unsubscribe from this group and all its topics, send an email to

>>>>>>>>>> swri-ros-pkg-d...@**googlegroups.com<swri-ros-pkg-d...@googlegroups.com>
>>>>>>>>>> <javascript:>.
>>>>>>>>>>
>>>>>>>>>> For more options, visit https://groups.google.com/**groups/opt_out<https://groups.google.com/groups/opt_out>

>>>>>>>>>> .
>>>>>>>>>>
>>>>>>>>>> --
>>>>>>>>> You received this message because you are subscribed to the Google
>>>>>>>>>
>>>>>>>> Groups
>>>>>>
>>>>>>> "swri-ros-pkg-dev" group.
>>>>>>>>> To unsubscribe from this group and stop receiving emails from it,
>>>>>>>>>
>>>>>>>> send
>>>>
>>>>> an
>>>>>>

>>>>>>> email to swri-ros-pkg-d...@**googlegroups.com<swri-ros-pkg-d...@googlegroups.com>
>>>>>>>>> <javascript:>.
>>>>>>>>> For more options, visit https://groups.google.com/**groups/opt_out<https://groups.google.com/groups/opt_out>
>>>>>>>>> .
>>>>>>>>>
>>>>>>>>
>>>>
>>>>
>>>
>> --

>> You received this message because you are subscribed to a topic in the
>> Google Groups "swri-ros-pkg-dev" group.

>> To unsubscribe from this topic, visit https://groups.google.com/d/**
>> topic/swri-ros-pkg-dev/uDu7O-**18Z1Q/unsubscribe<https://groups.google.com/d/topic/swri-ros-pkg-dev/uDu7O-18Z1Q/unsubscribe>

>> .
>> To unsubscribe from this group and all its topics, send an email to

>> swri-ros-pkg-dev+unsubscribe@**googlegroups.com<swri-ros-pkg-dev%2Bunsu...@googlegroups.com>
>> .
>>
>> For more options, visit https://groups.google.com/**groups/opt_out<https://groups.google.com/groups/opt_out>
>> .
>>
>

[ramo...@gmail.com]

Hi, I follow all the steps in the tutorial but when I launch the robot package (I use 2400 model) rviz doesn't recognize the kinematic chain (if I type rosrun tf view_frames only appears link_6  -->  tool0).

Any idea?

[Shaun Edwards]

What version of the ABB packages are you using?  Is it source or debian....hydro or groovy?

On Fri, May 23, 2014 at 3:32 AM, <ramo...@gmail.com> wrote:

Hi, I follow all the steps in the tutorial but when I launch the robot package (I use 2400 model) rviz doesn't recognize the kinematic chain (if I type rosrun tf view_frames only appears link_6  -->  tool0).

Any idea?

--
You received this message because you are subscribed to the Google Groups "swri-ros-pkg-dev" group.

To unsubscribe from this group and stop receiving emails from it, send an email to swri-ros-pkg-d...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[G.A. vd. Hoorn - 3ME]

See also [1].


Gijs

[1] http://answers.ros.org/question/167801/abb-robotstudio