Add doc page about joint kinematics (backport #1497)

[mergify]

I summarized some results of discussions about mimic joints and how I understand that.
Note, this is different than implemented currently with

ros2_control/hardware_interface/src/mock_components/generic_system.cpp

Lines 568 to 576 in 16fbde3

	for (const auto & mimic_joint : info_.mimic_joints)
	{
	for (auto i = 0u; i < joint_states_.size(); ++i)
	{
	joint_states_[i][mimic_joint.joint_index] =
	mimic_joint.offset +
	mimic_joint.multiplier * joint_states_[i][mimic_joint.mimicked_joint_index];
	}
	}

This came from the time where we defined the mimic joints in the ros2_control tag. IMHO, it does not make sense now using the URDF-specification, where the behavior of the mimic joint changes by switching the command interface from position to velocity, because we use the offset and calculate
velocity = multiplier * other_joint_velocity + offset

Also related to ros-controls/gazebo_ros2_control#297 and ros-controls/gz_ros2_control#276

I also included transmission interfaces, to give an overview what they are and what they aren't (they were not mentioned once in the docs).

---

This is an automatic backport of pull request #1497 done by [Mergify](https://mergify.com).