Re: [CoMPS-users] DoGeneralIK ignores repeating indices for active DOF indices.

[Dmitry B. <dbe...@cs...>]

Hi Ben,

This is interesting, I've never tried putting in duplicate DOF indices. I
think in general you may not want to temp fate by duplicating indices in
openrave's activedofs. For instance, if you have a vector of joint values
and you put different joint values for the same DOF, I'm not sure what
openrave would do. If GeneralIK is doing that the results could be pretty
ugly. I would suggest that you filter out the duplicate active DOF before
calling SetActiveDOFs anywhere in openrave (i.e. right after you get the
indices from the manipulators).

Dmitry



On Wed, Jul 9, 2014 at 1:40 PM, Bener Suay <ben...@gm...> wrote:

> Hi,
>
> my robot model has two 10 DOF manipulators and the first three joints of
> these manipulators are the same. To be more specific:
>
> leftArm indices:  0,1,2,3,4,5,6,7,8,9
> rightArm indices: 0,1,2,10,11,12,13,14,15,16
>
> In OpenRAVE (0.8.2 Stable Release), I do the following to activate my two
> manipulators so I can run DoGeneralIK for the end effectors:
>
> for( size_t m=0; m < manipIdx.size(); m++)
>     {
>       activedofs.insert( activedofs.end(),
>
> probot->GetManipulators()[manipIdx[m]]->GetArmIndices().begin(),
>
> probot->GetManipulators()[manipIdx[m]]->GetArmIndices().end()
>                  );
>     }
>
> The block above gives me the vector<int> activedofs as
>
> {0,1,2,3,4,5,6,7,8,9,0,1,2,10,11,12,13,14,15,16}
>
> And then I do
>
> probot->SetActiveDOFs(activedofs);
>
> as usual.
>
> To confirm that the indices are actually active I do:
>
> vector<int> dbug = probot->GetActiveDOFIndices();
> cout << "checking which indices are active: " << endl;
> // debug
> for( size_t i=0; i < dbug.size(); i++)
>   {
>      cout << dbug[i] << endl;
>    }
>
> And this prints out the same indices.
>
> So far so good...
>
> Now when I make a call to DoGeneralIK with this set of active indices, the
> joints that overlap between manipulators (namely 0, 1 and 2) behave as if
> they are not active (that is, they don't move).
>
> *However*, if I eliminate the repeating indices with the following two
> lines
>
> sort(activedofs.begin(), activedofs.end());
> activedofs.erase(unique(activedofs.begin(), activedofs.end()),
> activedofs.end());
>
> so that, my vector of activedofs is
>
> {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16}
>
> In this case DoGeneralIK behaves as expected.
>
> I'm wondering if this is a SetActiveDOFs issue, or if this is related to
> the way DoGeneralIK method handles GetActiveDOFIndices()... Curious.
>
> I'd be happy to contribute with the two-liner "remove repeating indices"
> block if that really is the appropriate / elegant solution for the issue.
>
> Thanks for your attention,
> Ben Suay
> wpi.edu/~benersuay
>
>
> ------------------------------------------------------------------------------
> Open source business process management suite built on Java and Eclipse
> Turn processes into business applications with Bonita BPM Community Edition
> Quickly connect people, data, and systems into organized workflows
> Winner of BOSSIE, CODIE, OW2 and Gartner awards
> http://p.sf.net/sfu/Bonitasoft
> _______________________________________________
> comps-users mailing list
> com...@li...
> https://lists.sourceforge.net/lists/listinfo/comps-users
>
>