Hi Ugo,

You can use KDL to do inverse kinematics using one of the inverse kinematics solvers in the package. These are numerical solvers that operate on a KDL chain. First use the kdl_parser package to create a KDL chain and then initialize an inverse kinematics solver for the chain. The solution tends to be dependent on the initial state you pass into the solver and can sometimes get stuck in local minima. Also, I am not sure about how well the solvers handle joint limits. I believe more advanced solvers are in the works but I will let the orocos folks comment on that.

Have a look at the api page here http://orocos.org/kdl/UserManual/kinematic_solvers for more examples.

The other option is to write your own solver (we have one for the PR2 that takes in a hint for one of the joint angles and then computes the other 6 analytically - www.ros.org/wiki/pr2_arm_kinematics)

Sachin

On Wed, Apr 7, 2010 at 3:28 AM, Ugo Cupcic <ugo@shadowrobot.com> wrote:

Hi all,

Is it possible to use robot_state_publisher or some other ROS stack to
do reverse kinematic on my urdf model?
Or do I need to implement my own using the "Coding a realtime Cartesian
controller with KDL" tutorial ?

Cheers,

Ugo

--
Ugo Cupcic | Shadow Robot Company | ugo@shadowrobot.com
Software Engineer 251 Liverpool Road
need a Hand? London N1 1LX | +44 20 7700 2487
http://www.shadowrobot.com/hand/ @shadowrobot

_______________________________________________
ros-users mailing list
ros-users@code.ros.org
https://code.ros.org/mailman/listinfo/ros-users

--
Sachin Chitta
Research Scientist
Willow Garage