Getting started with constraint_ik

[mlautman]

I'm working on a project using ros kinetics that would benefit greatly from a planner capable of object avoidance. I'm not entirely sure where to start. As a first test, I am trying to modify the stomp_test_kr210_moveit_config demo.launch to use kinematics_solver: constrained_ik/ConstrainedIKPlugin as described here

Doing so results in rviz crashing with this error message:

ros.constrained_ik: Unable to find ros parameter: constrained_ik_solver/manipulator/constraints
ros.constrained_ik: BREAKPOINT HIT
	file = /home/mike/ws_industrial/src/industrial_moveit/constrained_ik/src/constrained_ik.cpp
	line=57

[rviz_ubuntu_64011_6174358474395917019-5] process has died [pid 64058, exit code -5, cmd /opt/ros/kinetic/lib/rviz/rviz -d 

This makes me think that I need to upload the constraints to the param server before this will work.
I'm not entirely sure of the correct way to go about doing this because running the examples in the constrained_ik/examples directory (both as roslaunch constrained_ik goal_tool_position_example.launch and roslaunch constrained_ik launch_example.launch example:=goal_tool_position_example use_yaml:=true) result in the same error.

I've also tried following the other available tutorial on using constraint_ik but it isn't super clear how to bridge the gap between groovy and kinetic.

Thanks in advance!

[Levi-Armstrong]

Hello,
The links you provided are outdated which is why things were probably are not working. Here is the updated documentation. I recommend taking a look at the examples. There is a yaml section outline what is required for everything to work. Though the one example you mention should work so I am going to test it and see if I get the same results.

[Levi-Armstrong]

So it works if you set use_yaml:=false but not when set to true. I am working on a fix now.

[Levi-Armstrong]

I just created PR #61 to fix the examples when loading from yaml. Now when using the planner plugin with moveit you must replace the example_yaml with constrained_ik_solver.