[Robotics] COLLADA 1.5 armature support?

Herman Bruyninckx Herman.Bruyninckx at mech.kuleuven.be
Wed Feb 6 23:37:46 CET 2013


On Wed, 6 Feb 2013, William Hilton wrote:

> If I understand you right, then there are at least two types of joints,
> animation skeleton joints and kinematic model joints. Can Blender support
> all the features of kinematic models, such as limits?

It is better to put this kind of functionality into the MORSE extension to
Blender: MORSE is the right place to put all _robotics_-domain knowledge
and code.

> If not, perhaps it
> would make sense to prompt the user to ask if they want to convert the
> kinematic models to animation skeletons, at the loss of some kinematic
> model information. Does that sound like a good approach? Or should I try to
> modify my Collada file to change the joints to the type Blender supports?

My suggestion is to improve the way MORSE supports kinematic chains. _If_
that is necessary, since MORSE has already done some good steps in the
right direction, by extending the "Blender armature" concept with actuators
and sensors. (Nothing is perfect though, but that is more a problem of the
robotics domain overall and not of MORSE.)

> William Hilton

Herman

> MET-lab PhD Student, ECE Department, Drexel University
> Website: http://music.ece.drexel.edu/people/wmhilton
> Email: wmhilton at drexel.edu
> SMS: 513-368-5411
>
>
> On Wed, Feb 6, 2013 at 3:57 AM, Séverin Lemaignan <severin.lemaignan at laas.fr
>> wrote:
>
>> After a look to the Collada spec, what Blender supports are joints for
>> animation skeletons and skin (page 110 of Collada 1.5 spec). These joints
>> do not support limits for instance (only a transformation matrix).
>>
>> What OpenRAVE emits are (rightfully) kinematic models (chap. 10, p. 435 of
>> Collada spec). These are apparently not supported by the Collada Blender
>> importer. It should be however quite doable to add this feature to the
>> importer (since it has similar effects as the existing skeleton system).
>>
>> You may want to give it a look...
>>
>> Cheers,
>> Severin
>>
>>
>> On 02/06/2013 02:24 AM, William Hilton wrote:
>>
>>> Hi Severin,
>>> I started looking at the actual XML, and have decided COLLADA is more
>>> complicated than I would have thought! There may be multiple ways to
>>> define an armature, and maybe my .dae file doesn't do it in a way
>>> Blender understands. Honestly, I don't know if the problem lies in
>>> Blender's import. For all I know, it could be in how OpenRAVE exports
>>> the model to .dae.
>>>
>>> Here's the file I am trying to use. It's 6MB so I'll put it in Google
>>> Docs.
>>> https://docs.google.com/file/**d/0B7vT3fhK-**GBERGVDc0lWTVpBZnM/edit?usp=
>>> **sharing<https://docs.google.com/file/d/0B7vT3fhK-GBERGVDc0lWTVpBZnM/edit?usp=sharing>
>>> Will
>>>
>>> --
>>> William Hilton
>>> MET-lab PhD Student, ECE Department, Drexel University
>>> Website: http://music.ece.drexel.edu/**people/wmhilton<http://music.ece.drexel.edu/people/wmhilton>
>>> Email: wmhilton at drexel.edu <mailto:wmhilton at drexel.edu>
>>>
>>> SMS: 513-368-5411
>>>
>>>
>>> On Tue, Feb 5, 2013 at 3:24 PM, Séverin Lemaignan
>>> <severin.lemaignan at laas.fr <mailto:severin.lemaignan@**laas.fr<severin.lemaignan at laas.fr>>>
>>> wrote:
>>>
>>>     After some collada + armatures testing (namely, exporting
>>>     kuka_arm.blend to collada and re-importing, and importing this
>>>     collada file [1], with blender 2.65a), I get mixed results:
>>>     the Kuka model is garbage when I re-import it, while the collada
>>>     sample [1] looks fine.
>>>
>>>     So it is possible the Blender Collada *exporter*  does not work well
>>>     with armatures. Can you share a collada model you would like to
>>> import?
>>>
>>>     Cheers,
>>>     Severin
>>>
>>>     [1]
>>>     http://anima-animation-sample.**__googlecode.com/svn-history/_**
>>> _r11/trunk/Models/cubes2.dae<http://googlecode.com/svn-history/__r11/trunk/Models/cubes2.dae>
>>>
>>>     <http://anima-animation-**sample.googlecode.com/svn-**
>>> history/r11/trunk/Models/**cubes2.dae<http://anima-animation-sample.googlecode.com/svn-history/r11/trunk/Models/cubes2.dae>
>>>>
>>>
>>>
>>>     On 05/02/2013 20:15, William Hilton wrote:
>>>
>>>         Yes, I should have mentioned I am using the latest version of
>>>         Blender
>>>         (2.65a).
>>>
>>>
>>>
>>>     --
>>>               Séverin Lemaignan - severin.lemaignan at laas.fr
>>>     <mailto:severin.lemaignan@**laas.fr <severin.lemaignan at laas.fr>>
>>>
>>>
>>>       [00]    Post-doctoral fellow in Cognitive Robotics
>>>     /|__|\   LAAS-CNRS - RIS group
>>>        '' +33561337844 <tel:%2B33561337844>
>>>     http://www.laas.fr/~slemaign
>>>
>>>
>>>
>> --
>>           Séverin Lemaignan - severin.lemaignan at laas.fr
>>   [00]    Post-doctoral fellow on Cognitive Robotics
>>  /|__|\   LAAS-CNRS - RIS group
>>    ''     +33561337844
>>           http://www.laas.fr/~slemaign
>>
>

-- 
--
   KU Leuven, Mechanical Engineering, Robotics Research Group
     <http://people.mech.kuleuven.be/~bruyninc> Tel: +32 16 328056
   Vice-President Research euRobotics <http://www.eu-robotics.net>
   Open RObot COntrol Software <http://www.orocos.org>
   Associate Editor JOSER <http://www.joser.org>, IJRR <http://www.ijrr.org>


More information about the Robotics mailing list