[ros2_control] What are the advantages/limitations of using ActuatorInterface over SystemInterface?
The ros2_control documentation suggests that ActuatorInterface is a sub-case of SystemInterface for when there is only 1 joint/DoF. However, nothing is stopping me from creating an ActuatorInterface with multiple joints and multiple state/command interfaces for each joint (as you would have for a SystemInterface). I took a SystemInterface I implemented and changed it to an ActuatorInterface (and updated the urdf to use "actuator" instead of "system") and everything still works as it should.
So are Systems and Actuators exactly the same / interchangeable and the reason to differentiate between them is for naming convention? Or are there actually advantages/limitations to using System vs Actuator aside from naming convention?
Let's give the
ros2_control
maintainers some time to answer ros-controls/ros2_control#939 which you posted less than a day ago.Okay. I re-posted it here because I realized (after posting the github issue) that the ros2_control documentation (https://control.ros.org/master/index....) says to post questions on ROS answers.
If ROS Answers is the correct place for the question and it can't be on ROS Answers with the duplicate question open on github, I can close the issue on github. Let me know.
Apologies for the duplicate question. I was trying to move it to the correct forum after realizing github issues wasn't the recommended place according to ros2_control documentation.
If you want to keep this one we can re-open it. I'd suggest to close your Github issue then (perhaps after posting a comment there linking to your ROS Answers post here).
I just closed the Github issue. Please re-open this one. Thanks!