I'm working on a Schunk LWA 4P and I want to use the IKFast for tasks of teleoperation. The model provided by the GitHub packages "schunk_robots" and "schunk_modular_robotics" are not very accurate in the dimensions and the use of the axis in the base, so I created my own one following the specs sheet provided with the robot.
The problem that I found when I try to apply the code generator for this model is that it says that it doesn't find three consecutive intersecting axis, when indeed there are. The robot can be easily decoupled in two parts for orientation and position with the 3 first and 3 following joints.
There is another post in this forum where a guy had a similar problem, and he provided the collada file, which had been supposedly extracted from the previously mentioned packages. With the file provided by this person it generates the code well. But if you convert to collada the files provided in those packages the generator fails for the same reason (non intersecting axis). Here I leave the link to that conversation:
Another question that I have is that inside the collada file, which reference is the one taken into account when the parameters "--baselink=" and "--eelink=" have to be specified to create the code. Is the position in the <library_kinematics_models > or the node0, node1, etc, defined at the beginning in the <library_visual_scenes >? Because these last ones normally are taken in alphabetical order instead of the kinematic tree order.
Furthermore I attach a simplified version of the collada file of the created robot in case you can tell me the error revising it. Is the one used to pass as parameter to the python script.