ikfast solver does not find solutions

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

ikfast solver does not find solutions

andreas
Hi all,

I've tried several methods of generating an ikfast solver from the attached dae/xml file, such as:

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w1 --freeinc=0.01

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w0 --freeinc=0.01

However, both methods give the following output:

openravepy.databases.inversekinematics: testik, success rate: 0.000000, wrong solutions: 0.000000, no solutions: 1.000000, missing solution: 0.000000

This means that NO solution was found for any of the 100 ik tests.

-- Andreas




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

_______________________________________________
Openrave-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/openrave-users

baxter.xml (234 bytes) Download Attachment
baxter_arm.right.rounded.dae (106K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: ikfast solver does not find solutions

baxelrod
I have never had much luck with the iktests.  They always seem to fail.  Even though the files generated by ikfast work fine.
-Ben

-----Original Message-----
From: Andreas ten Pas [mailto:[hidden email]]
Sent: Friday, July 31, 2015 8:35 AM
To: openrave-users
Subject: [OpenRAVE-users] ikfast solver does not find solutions

Hi all,

I've tried several methods of generating an ikfast solver from the attached dae/xml file, such as:

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w1 --freeinc=0.01

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w0 --freeinc=0.01

However, both methods give the following output:

openravepy.databases.inversekinematics: testik, success rate: 0.000000, wrong solutions: 0.000000, no solutions: 1.000000, missing solution: 0.000000

This means that NO solution was found for any of the 100 ik tests.

-- Andreas



------------------------------------------------------------------------------
_______________________________________________
Openrave-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/openrave-users
Reply | Threaded
Open this post in threaded view
|

Re: ikfast solver does not find solutions

andreas
Thanks for your input.

The generated ik solver does not give an ik solution in python either. I attached two python codes, one that uses databases inversekinematics and one that does not.

-- Andreas

----- Original Message -----
From: "Axelrod, Ben" <[hidden email]>
To: "Andreas ten Pas" <[hidden email]>, "openrave-users" <[hidden email]>
Sent: Friday, July 31, 2015 9:22:56 AM
Subject: RE: ikfast solver does not find solutions

I have never had much luck with the iktests.  They always seem to fail.  Even though the files generated by ikfast work fine.
-Ben

-----Original Message-----
From: Andreas ten Pas [mailto:[hidden email]]
Sent: Friday, July 31, 2015 8:35 AM
To: openrave-users
Subject: [OpenRAVE-users] ikfast solver does not find solutions

Hi all,

I've tried several methods of generating an ikfast solver from the attached dae/xml file, such as:

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w1 --freeinc=0.01

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w0 --freeinc=0.01

However, both methods give the following output:

openravepy.databases.inversekinematics: testik, success rate: 0.000000, wrong solutions: 0.000000, no solutions: 1.000000, missing solution: 0.000000

This means that NO solution was found for any of the 100 ik tests.

-- Andreas
------------------------------------------------------------------------------

_______________________________________________
Openrave-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/openrave-users

baxter_ik.xml (648 bytes) Download Attachment
iktest2.py (1K) Download Attachment
iktest3.py (1K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: ikfast solver does not find solutions

Rosen Diankov
Administrator
Daer Andreas,

Can you try the latest ikfast (from openrave master branch)?
There were several fixes in the past weeks that resolve problems like the ones you are seeing.

Sincerely,
Rosen,


2015-08-01 0:53 GMT+09:00 Andreas ten Pas <[hidden email]>:
Thanks for your input.

The generated ik solver does not give an ik solution in python either. I attached two python codes, one that uses databases inversekinematics and one that does not.

-- Andreas

----- Original Message -----
From: "Axelrod, Ben" <[hidden email]>
To: "Andreas ten Pas" <[hidden email]>, "openrave-users" <[hidden email]>
Sent: Friday, July 31, 2015 9:22:56 AM
Subject: RE: ikfast solver does not find solutions

I have never had much luck with the iktests.  They always seem to fail.  Even though the files generated by ikfast work fine.
-Ben

-----Original Message-----
From: Andreas ten Pas [mailto:[hidden email]]
Sent: Friday, July 31, 2015 8:35 AM
To: openrave-users
Subject: [OpenRAVE-users] ikfast solver does not find solutions

Hi all,

I've tried several methods of generating an ikfast solver from the attached dae/xml file, such as:

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w1 --freeinc=0.01

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w0 --freeinc=0.01

However, both methods give the following output:

openravepy.databases.inversekinematics: testik, success rate: 0.000000, wrong solutions: 0.000000, no solutions: 1.000000, missing solution: 0.000000

This means that NO solution was found for any of the 100 ik tests.

-- Andreas

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

_______________________________________________
Openrave-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/openrave-users



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

_______________________________________________
Openrave-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/openrave-users
Reply | Threaded
Open this post in threaded view
|

Re: ikfast solver does not find solutions

baxelrod
Bugs aside, the difference between these setups is which joint is free.  I have an open question regarding this topic on Robotics.StackExchange:
http://robotics.stackexchange.com/questions/7786/which-joints-to-discretize-for-ik

Put simply, is it better to have free joints closer to the base or end-effector?  It seems to me that it is probably better to have joints closer to the end-effector as free joints to avoid lack of solutions with large discretizations.  However, ikfast seems to pick joints closer to the base...

Thanks,
Ben

Ben Axelrod

Sr. Robotics Software Engineer

Technology Organization


iRobot

8 Crosby Drive 8-1

Bedford, MA 01730

t: 781.430.3315

[hidden email]


From: [hidden email] [[hidden email]] on behalf of Rosen Diankov [[hidden email]]
Sent: Saturday, August 01, 2015 2:39 AM
To: Andreas ten Pas
Cc: Axelrod, Ben; openrave-users
Subject: Re: [OpenRAVE-users] ikfast solver does not find solutions

Daer Andreas,

Can you try the latest ikfast (from openrave master branch)?
There were several fixes in the past weeks that resolve problems like the ones you are seeing.

Sincerely,
Rosen,


2015-08-01 0:53 GMT+09:00 Andreas ten Pas <[hidden email]>:
Thanks for your input.

The generated ik solver does not give an ik solution in python either. I attached two python codes, one that uses databases inversekinematics and one that does not.

-- Andreas

----- Original Message -----
From: "Axelrod, Ben" <[hidden email]>
To: "Andreas ten Pas" <[hidden email]>, "openrave-users" <[hidden email]>
Sent: Friday, July 31, 2015 9:22:56 AM
Subject: RE: ikfast solver does not find solutions

I have never had much luck with the iktests.  They always seem to fail.  Even though the files generated by ikfast work fine.
-Ben

-----Original Message-----
From: Andreas ten Pas [mailto:[hidden email]]
Sent: Friday, July 31, 2015 8:35 AM
To: openrave-users
Subject: [OpenRAVE-users] ikfast solver does not find solutions

Hi all,

I've tried several methods of generating an ikfast solver from the attached dae/xml file, such as:

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w1 --freeinc=0.01

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w0 --freeinc=0.01

However, both methods give the following output:

openravepy.databases.inversekinematics: testik, success rate: 0.000000, wrong solutions: 0.000000, no solutions: 1.000000, missing solution: 0.000000

This means that NO solution was found for any of the 100 ik tests.

-- Andreas

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

_______________________________________________
Openrave-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/openrave-users



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

_______________________________________________
Openrave-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/openrave-users
Reply | Threaded
Open this post in threaded view
|

Re: ikfast solver does not find solutions

andreas
In reply to this post by Rosen Diankov
Hello Rosen,

thanks!! With the newest version, generating an ikfast solver worked fine.

-- Andreas

----- Original Message -----
From: "Rosen Diankov" <[hidden email]>
To: "Andreas ten Pas" <[hidden email]>
Cc: "Axelrod, Ben" <[hidden email]>, "openrave-users" <[hidden email]>
Sent: Saturday, August 1, 2015 2:39:08 AM
Subject: Re: [OpenRAVE-users] ikfast solver does not find solutions

Daer Andreas,

Can you try the latest ikfast (from openrave master branch)?
There were several fixes in the past weeks that resolve problems like the
ones you are seeing.

Sincerely,
Rosen,


2015-08-01 0:53 GMT+09:00 Andreas ten Pas <[hidden email]>:

> Thanks for your input.
>
> The generated ik solver does not give an ik solution in python either. I
> attached two python codes, one that uses databases inversekinematics and
> one that does not.
>
> -- Andreas
>
> ----- Original Message -----
> From: "Axelrod, Ben" <[hidden email]>
> To: "Andreas ten Pas" <[hidden email]>, "openrave-users" <
> [hidden email]>
> Sent: Friday, July 31, 2015 9:22:56 AM
> Subject: RE: ikfast solver does not find solutions
>
> I have never had much luck with the iktests.  They always seem to fail.
> Even though the files generated by ikfast work fine.
> -Ben
>
> -----Original Message-----
> From: Andreas ten Pas [mailto:[hidden email]]
> Sent: Friday, July 31, 2015 8:35 AM
> To: openrave-users
> Subject: [OpenRAVE-users] ikfast solver does not find solutions
>
> Hi all,
>
> I've tried several methods of generating an ikfast solver from the
> attached dae/xml file, such as:
>
> openrave.py --database inversekinematics --robot=baxter.xml
> --iktype=transform6d --iktests=100 --freejoint=right_w1 --freeinc=0.01
>
> openrave.py --database inversekinematics --robot=baxter.xml
> --iktype=transform6d --iktests=100 --freejoint=right_w0 --freeinc=0.01
>
> However, both methods give the following output:
>
> openravepy.databases.inversekinematics: testik, success rate: 0.000000,
> wrong solutions: 0.000000, no solutions: 1.000000, missing solution:
> 0.000000
>
> This means that NO solution was found for any of the 100 ik tests.
>
> -- Andreas
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Openrave-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/openrave-users
>
>

------------------------------------------------------------------------------
_______________________________________________
Openrave-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/openrave-users
Reply | Threaded
Open this post in threaded view
|

Re: ikfast solver does not find solutions

Rosen Diankov
Administrator
In reply to this post by baxelrod
Dear Ben,

That's an excellent question, and actually very difficult to answer. The general rule of thumb is that the closer it is to the end effector the better, but this is not always the case. For example, there's only one set of 3 intersecting axes and that's the final 3 joints, then choosing a free joint at the end will mean the resulting IK solver cannot exploit the 3 intersecting axis property to simplify the IK.

The only way I've found to truly get the best free joint is to solve for all of them and see which one is better ;0)


2015-08-01 22:19 GMT+09:00 Axelrod, Ben <[hidden email]>:
Bugs aside, the difference between these setups is which joint is free.  I have an open question regarding this topic on Robotics.StackExchange:
http://robotics.stackexchange.com/questions/7786/which-joints-to-discretize-for-ik

Put simply, is it better to have free joints closer to the base or end-effector?  It seems to me that it is probably better to have joints closer to the end-effector as free joints to avoid lack of solutions with large discretizations.  However, ikfast seems to pick joints closer to the base...

Thanks,
Ben

Ben Axelrod

Sr. Robotics Software Engineer

Technology Organization


iRobot

8 Crosby Drive 8-1

Bedford, MA 01730

t: 781.430.3315

[hidden email]


From: [hidden email] [[hidden email]] on behalf of Rosen Diankov [[hidden email]]
Sent: Saturday, August 01, 2015 2:39 AM
To: Andreas ten Pas
Cc: Axelrod, Ben; openrave-users
Subject: Re: [OpenRAVE-users] ikfast solver does not find solutions

Daer Andreas,

Can you try the latest ikfast (from openrave master branch)?
There were several fixes in the past weeks that resolve problems like the ones you are seeing.

Sincerely,
Rosen,


2015-08-01 0:53 GMT+09:00 Andreas ten Pas <[hidden email]>:
Thanks for your input.

The generated ik solver does not give an ik solution in python either. I attached two python codes, one that uses databases inversekinematics and one that does not.

-- Andreas

----- Original Message -----
From: "Axelrod, Ben" <[hidden email]>
To: "Andreas ten Pas" <[hidden email]>, "openrave-users" <[hidden email]>
Sent: Friday, July 31, 2015 9:22:56 AM
Subject: RE: ikfast solver does not find solutions

I have never had much luck with the iktests.  They always seem to fail.  Even though the files generated by ikfast work fine.
-Ben

-----Original Message-----
From: Andreas ten Pas [mailto:[hidden email]]
Sent: Friday, July 31, 2015 8:35 AM
To: openrave-users
Subject: [OpenRAVE-users] ikfast solver does not find solutions

Hi all,

I've tried several methods of generating an ikfast solver from the attached dae/xml file, such as:

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w1 --freeinc=0.01

openrave.py --database inversekinematics --robot=baxter.xml --iktype=transform6d --iktests=100 --freejoint=right_w0 --freeinc=0.01

However, both methods give the following output:

openravepy.databases.inversekinematics: testik, success rate: 0.000000, wrong solutions: 0.000000, no solutions: 1.000000, missing solution: 0.000000

This means that NO solution was found for any of the 100 ik tests.

-- Andreas

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

_______________________________________________
Openrave-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/openrave-users




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

_______________________________________________
Openrave-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/openrave-users