Hi, I am not an expert, but this is what I think from my experience so far:
1) In rviz I think you can only set one goal at a time. If your robot is navigating and you use the 2D Navigation button again, your robot will forget what it was doing and navigate to the new goal.
2) I have never tried this, but I imagine if the initial pose is far off it would be difficult for the robot to localize itself.
3) By autonomous navigation, do you mean this tutorial?
http://wiki.ros.org/turtlebot_navigation/Tutorials/Autonomously%20navigate%20in%20a%20known%20map
The tutorial includes this paragraph:
Teleoperation
The teleoperation can be run
simultaneously with the navigation
stack. It will override the autonomous
behavior if commands are being sent.
It is often a good idea to teleoperate
the robot after seeding the
localization to make sure it converges
to a good estimate of the position.
So I believe you should be able to run both simultaneously. If you're doing the autonomous tutorial you should already have your robot booted up, so I would think all you need to do is open a new terminal (Ctrl+Alt+T) and type in roslaunch turtlebot_teleop keyboard_teleop.launch
(I don't know what robot you're using. I am using a Turtlebot, so that is the teleop command for it)
4)
For local_costmap_params,
global_frame: /odom
robot_base_frame: /base_footprint
For global_costmap_params,
global_frame: /map
robot_base_frame: /base_footprint
I hope this helps!
