Re: Linear Cartesian movements
by ipa-lth » Wed 22. Oct 2014, 13:28
Dear AIS,
the node providing /simulation_interface_ur5/cartesian_position has been used for some basic proof of concept and is just not capable of performing what you want.
If I am not mistaken, the program interpolates exactly 10 waypoints between the actual pose and the target pose, uses the KDL inverse kinematic solver to transform the waypoints to joint configurations and send them to the joint controller. Thereby effectivitly performing a PTP movement with 10 intermediate points on a straight line. The time between every configuration is hard coded, and that is why long pathes are very course while short pathes take very long.
In a nutshell "simulation_interface_ur5" is just not powerful, robust or well-tested enough.
Probably the best possibilities for you are to use either Move-it or implement your own solution.
Sorry for the bad news,
kind regards,
Lorenz
___________________________________________________________________
Fraunhofer Institut for Manufacturing Engineering and Automation (IPA)
Department for Robot and Assistive Systems
EuRoC Challenge 1 - Support