[jog_arm] Fix valid cmd flags

[AndyZe]

Switching back and forth between joint commands and Cartesian commands could be buggy sometimes. The change at line 203 of jog_calcs.cpp fixes that.

Some flags were renamed and their logic was flipped to help readability

[AndyZe]

@JStech @henningkayser

[codecov-io]

Codecov Report

Merging #2013 into master will increase coverage by 0.06%.
The diff coverage is 67.85%.

@@            Coverage Diff             @@
##           master    #2013      +/-   ##
==========================================
+ Coverage   51.82%   51.88%   +0.06%     
==========================================
  Files         319      317       -2     
  Lines       24959    24618     -341     
==========================================
- Hits        12934    12773     -161     
+ Misses      12025    11845     -180     

Impacted Files	Coverage Δ
...veit_jog_arm/include/moveit_jog_arm/jog_arm_data.h	100.00% <ø> (ø)
...erimental/moveit_jog_arm/src/jog_cpp_interface.cpp	11.49% <0.00%> (ø)
...veit_experimental/moveit_jog_arm/src/jog_calcs.cpp	75.31% <100.00%> (+0.63%)	⬆️
...erimental/moveit_jog_arm/src/jog_ros_interface.cpp	84.81% <100.00%> (ø)
.../include/moveit/py_bindings_tools/py_conversions.h	0.00% <0.00%> (-86.67%)	⬇️
...dl_kinematics_plugin/src/kdl_kinematics_plugin.cpp	75.33% <0.00%> (-2.25%)	⬇️
...ls/include/moveit/py_bindings_tools/gil_releaser.h
...s/include/moveit/py_bindings_tools/serialize_msg.h
...anning_scene_monitor/src/current_state_monitor.cpp	52.82% <0.00%> (+1.53%)	⬆️
... and 2 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d989220...83b46b7. Read the comment docs.

[henningkayser]

@AndyZe Couldn't this potentially lead to sudden jerks in the motion when switching from one mode to another? In general, I think it might make sense to make this a little more explicit, i.e. setting the mode via CPP-interface, or only switching mode after repeated commands via ROS-interface

[henningkayser]

Any straight-forward way to test this? Switching modes is more of an edge case for now I assume, but would it be possible to mix joint and cartesian commands without much delay?

[AndyZe]

@AndyZe Couldn't this potentially lead to sudden jerks in the motion when switching from one mode to another? In general, I think it might make sense to make this a little more explicit, i.e. setting the mode via CPP-interface, or only switching mode after repeated commands via ROS-interface

Well, the user can already request abrupt changes in motion. In Cartesian mode, the commands get filtered, but there isn't a filter in joint mode. A joint command filter might be something to add for the future.

Yeah, an explicit mode is a good idea, although a pain to implement and test for C++ and the ROS interface.

Edit - I looked through the code to see how the lowpass filters get applied. They are used in both Cartesian and joint control mode, so commands from the user should get smoothed regardless. So, I don't think an explicit mode is necessary. It's nice to keep the jogger agnostic to the source of incoming commands.

I'll make some plots to verify the smoothing actually happens when transitioning from Cartesian -> joint control.

[JStech]

LGTM. Thanks for tracking this down

[AndyZe]

Here are some plots of transitions from joint to Cartesian commands in opposite directions. And Cartesian commands to joint commands. Looks pretty smooth to me, I don't think more work on that is necessary.

[henningkayser]

@AndyZe great work! Tested this and didn't find any issues with it.