comps-users Mailing List for Constrained Manipulation Planning Suite

openrave plugins for constrained manipulation

Brought to you by: berenson, cdellin, vandeweg

comps-users — Questions and comments about CoMPS

You can subscribe to this list here.

2012	_Jan	_Feb	_Mar	_Apr	_May	_Jun	_Jul	_Aug (14)	_Sep	_Oct	_Nov	_Dec
2014	_Jan	_Feb	_Mar	_Apr	_May	_Jun	_Jul (3)	_Aug	_Sep	_Oct	_Nov	_Dec

Flat | Threaded

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

From: Bener S. <ben...@gm...> - 2014-07-09 17:55:42

OK, so then I'll just keep my code as it is (the way I sort and erase
duplicates).

I would think that SetActiveDOFs would do that for me internally for its
own error handling but apparently it doesn't, which is fine.

Thanks for the advice!

Ben



2014-07-09 10:51 GMT-07:00 Dmitry Berenson <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
>>
>>
>

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

From: Dmitry B. <dbe...@cs...> - 2014-07-09 17:52:23

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
>
>

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

From: Bener S. <ben...@gm...> - 2014-07-09 17:40:27

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