feat(hcm): setup callback groups for high hz subscriptions

[texhnolyze]

using a MutuallyExclusiveCallbackGroup for each one, so that they do not block each other but each subscription's callback is still called in order of received messages.

See: https://docs.ros.org/en/rolling/How-To-Guides/Using-callback-groups.html

Fixes: #313

Checklist

• Run colcon build
• Write documentation
• Test on your machine
• Test on the robot
• Create issues for future work
• Triage this PR and label it

[texhnolyze]

I tested in simulation and on robot and everything seems to work.
@Flova do you have an idea how I could benchmark if this lead to a performance improvement?

[timonegk]

Code-wise this looks fine but I cannot judge the performance increase.

[texhnolyze]

I took a look at the EventsExecutor implementation and determined, that it is single threaded and utilizing a single FIFO queue, no matter the callback groups.
To actually utilize parallel execution, we would need to either need to utilize an MultiThreadedExecutor with a multiple threads or assign different callback groups to separate instances of EventsExecutor utilizing the ->add_callback_group method.

In the way it was implemented right now, it will most likely not have any effect on the execution of callbacks in the HCM, but will also not break anything.

Until some clarification and benchmarking, I would put this on hold.
@Flova did we even still have issues with speed of joint_states and imu/data in the HCM in the last months?

[Flova]

I don't think so tbh. as discussed in person. (Just writing it here as a protocol for the future.