I'm currently working with the ChainIkSolverVel_pinv-class to calculate the inverse velocity for a robotic arm with 7 joints. After executing ChainIkSolverVel_pinv::CartToJnt I have only 6 joint velocities. The velocity of joint 7 is always 0. I've tried this for many end effector velocities and joint values (= input parameters of CartToJnt) but the value of the 7th joint stays 0.
So, I guess that CartToJnt calculate only 6 joints to avoid needless calculations of the other joints (= avoiding redundancy)?
Is that true?
Does ChainIkSolverVel_pinv create only 6 joint-velocity values?
Submitted by threelight on Mon, 2008-06-23 12:09 |
Does ChainIkSolverVel_pinv create only 6 joint-velocity values?
On Monday 23 June 2008 14:09:37 threelight wrote:
> I'm currently working with the ChainIkSolverVel_pinv-class to calculate the
> inverse velocity for a robotic arm with 7 joints. After executing
> ChainIkSolverVel_pinv::CartToJnt I have only 6 joint velocities. The
> velocity of joint 7 is always 0. I've tried this for many end effector
> velocities and joint values (= input parameters of CartToJnt) but the value
> of the 7th joint stays 0. So, I guess that CartToJnt calculate only 6
> joints to avoid needless calculations of the other joints (= avoiding
> redundancy)? Is that true?
No, the algorithm is based on a pseudo-inverse of the robot jacobian, it
should distribute the velocity equaly over all joints if possible.
What version of KDL are you using?
There was bug in the pseudo-inverse code, it should be fixed in kdl-trunk. I
tested the algorithm on a chain with 8 joints, all jointvelocities were
different from zero.
Ruben
I'm using version 1.0 So,
I'm using version 1.0
So, I'm using the wrong version?
And the bug was responsible for returning zero-values?
I'm using version 1.0
On Monday 23 June 2008 15:38:48 threelight wrote:,orocos-users-request [..] ...?subject=unsubscribe>orocos-users [..] ...>orocos-users-request [..] ...?subject=help>,orocos-users-request [..] ...?subject=subscribe>
> So,
> X-BeenThere: orocos-users [..] ...
> X-Mailman-Version: 2.1.5
> Precedence: list
> List-Id: Orocos Users
> List-Unsubscribe:
>
>
> List-Archive:
> List-Post:
> List-Help:
>
> List-Subscribe:
>
>
> Sender: orocos-users-bounces [..] ...
> Errors-To: orocos-users-bounces [..] ...
>
>
> I'm using version 1.0
>
> So, I'm using the wrong version?
The best version to use is still the one on svn trunk ;)
But maybe i should prepare a new release to get rid of this kind of problems.
> And the bug was responsible for returning zero-values?
Yes. Definitely.
Ruben
thanks!
thanks!