Trac-IK plugin

[rjoomen]

I've created a tesseract_kinematics plugin for the Trac-IK inverse kinematics solver. Would you be interested in including this in the main repo?

Problem is that it currently depends on a non-ROS-agnostic ROS2 fork. The main repo now also is updated for ROS2, but is even less ROS-agnostic (needs a node as parameter).

If we'd want to include this, probably forking it under tesseract_robotics and making the trac_ik_lib subproject ROS-agnostic would be the way to go. Afterwards we should try to get this restructured version PR'd to the original repo.

[Levi-Armstrong]

I am interested in getting track ik integrated. I would post the question on track ik to see if they would be interested in making track_ik_lib ROS-agnostic. If not, it looks like the track_ik_lib is stable so we could just copy the contents into tesseract_kinematics along with your plugin and make it an optional build.

[rjoomen]

I've added the plugin as a new project, pending the creation of a ROS-agnostic Trac-IK fork. Is this OK with you?

[gavanderhoorn]

I would suggest just use the upstream rolling-devel version. The fork you mention is outdated and no longer maintained/developed.

As to a ROS-agnostic version: there are already quite some forks which add that ability. See HIRO-group/trac_ik fi.

[rjoomen]

Upstream rolling-devel requires a rclcpp::Node as an argument, which Tesseract cannot provide. The fork I now use does not.

[rjoomen]

My PR to Trac_IK has been accepted: https://bitbucket.org/traclabs/trac_ik/pull-requests/40

tesseract_trac_ik now works with the official Trac_IK repo, instead of a fork. Trac_IK is still ROS-dependent, but for anyone using Tesseract with ROS, this is new IK plugin is a viable option.