Add pinocchio plugin

[christophfroehlich]

@justagist is the original author, but we discussed to move it into this repo to get proper maintenance

justagist/kinematics_interface_pinocchio#11 (comment)

[codecov]

Codecov Report

❌ Patch coverage is 85.37736% with 31 lines in your changes missing coverage. Please review.
✅ Project coverage is 86.09%. Comparing base (4a831a8) to head (61d8487).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
...e_pinocchio/src/kinematics_interface_pinocchio.cpp	75.78%	21 Missing and 10 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #199      +/-   ##
==========================================
- Coverage   86.64%   86.09%   -0.55%     
==========================================
  Files           4        6       +2     
  Lines         277      489     +212     
  Branches       60       80      +20     
==========================================
+ Hits          240      421     +181     
- Misses         21       42      +21     
- Partials       16       26      +10     

Flag	Coverage Δ
unittests	86.09% <85.37%> (-0.55%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
...terface_kdl/test/test_kinematics_interface_kdl.cpp	100.00% <100.00%> (ø)
...cchio/test/test_kinematics_interface_pinocchio.cpp	100.00% <100.00%> (ø)
...e_pinocchio/src/kinematics_interface_pinocchio.cpp	75.78% <75.78%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[christophfroehlich]

I'll draft it, because the plugin has some issues. The jacobian is not well-defined if the tool frame is not the end of the chain (it always has the size of DOF) -> this results in numerical flakiness and failing tests.

[qleonardolp]

I'll draft it, because the plugin has some issues. The jacobian is not well-defined if the tool frame is not the end of the chain (it always has the size of DOF) -> this results in numerical flakiness and failing tests.

I faced a similar issue writing my impedance controller. In my case, the jacobian was always considering the world frame as the base frame, thus a leg with 3-Dof was given a 3+6 dof Jacobian. My workaround was splitting the robot urdf between the part I need to pass to pinocchio, and the "whole" urdf to pass to Gazebo/Rviz2...

I'd love to her from you about this approach and how it could benefit the ros2_control framework with Pinocchio.

https://github.com/qleonardolp/ros2_impedance_controller
https://github.com/qleonardolp/ros2_descriptions

By the way, what do you guys think about a new impedance controller on ros2_controllers? 👀

[christophfroehlich]

Thanks for confirmation that there is no obvious fix ;) @saikishor mentioned that he might have an idea maybe?

[saikishor]

Thanks for confirmation that there is no obvious fix ;) @saikishor mentioned that he might have an idea maybe?

Yes, we can build a partial model out of the whole model and do the inverse kinematics etc