ROS Tutorials

• Log into one of the Turtlebot netbooks.
• Download the file cs354_ws.zip and unzip it in the robot home directory. (This zip file contains a properly configured catkin workspace containing a single ROS package named "labs". It was created by following the ROS tutorials on creating a workspace, creating a ROS package, and building a ROS package.)
• Complete ROS tutorials 12 and 13. You should be able to follow the steps of those tutorials exactly, except the workspace is named cs354_ws instead of catkin_ws, and the project is named labs instead of beginner_tutorials.

Applying Your Knowledge

• Connect and launch a Turtlebot as described in Lab 1. (Don't launch keyboard_teleop. You will be controlling the robot using a Python node.)

• Create a Python file named bumper.py that moves the robot forward until it bumps into something, and then stops. You can place your Python file in the same scripts directory you created in the tutorials above.

  Some advice: The organization of your python program will differ a bit from the talker and listener examples from the tutorial. In the tutorial, one node was a responsible for publishing messages, and a different node subscribed to those messages. In this case, a single node will act as both a publisher and a subscriber: your node must subscribe to a bumper topic and publish robot commands.

  This raises the question of how to share information between the sensor callback and the loop that is publishing control commands. A simple approach is to create a global variable that is modified by the callback and then accessed in the main loop. Something like this:

  #!/usr/bin/env python
  
  CALLBACK_DATA = None # Create a global variable
  
  def callback(data):
      global CALLBACK_DATA # 'global' keyword prevents Python from
                           # creating a new local (temporary) variable.
      CALLBACK_DATA = data 
  
  def main_loop():
      for i in range(10):
          print CALLBACK_DATA
  
  if __name__ == "__main__":
      callback("Howdy!")
      main_loop()
  

  A more elegant approach involves creating a Python class to represent your node. You can then use an instance variable to store data received from the callback:

  #!/usr/bin/env python
  
  class CallBacker(object):
      
      def __init__(self):
          self.callback_data = None
  
      def callback(self, data):
          self.callback_data = data
  
      def main_loop(self):
          for i in range(10):
              print self.callback_data
  
  if __name__ == "__main__":
      cb = CallBacker()
      cb.callback("Howdy!")
      cb.main_loop()
  
  

  You are welcome to use either approach for this assignment.

• Once you have finished your Python node, you should be able to test it by typing:

  rosrun labs bumper.py
  

  in a terminal window.

A Smarter Robot (If Time)

If you have extra time, improve your node so that when an obstacle is hit the robot:

• Moves backwards a few centimeters.
• Rotates in place 20^o to the left.
• Resumes forward progress.

Submitting

Submit your completed bumper.py file through Canvas. Each group should submit only one file. Comments in the file should include the names of all group members.