Homework 2 - More Robot Operating System (ROS)
Due at 11:00 pm on Thursday 09 February 2017
As discussed in class, we will be using the Robot Operating System (ROS)
in conjunction with TurtleBots in the course this semester. The point of
this homework is to take you through some tutorial exercises to help you
learn some of the basic features of ROS and TurtleBots. Note that, as with
Homework 1, you will not be programming any robots in this homework.
However, you will get to drive them around in simulation a bit!
Computer
Resources Note font> |
See the computer resources note from Homework 1.
|
The Assignment
Work through the ROS tutorials, completing the following exercises as
you go. As you complete these exercises and add items to your completed
assignment, number them to correspond to the numbering below as you did
with Homework 1. Note that not all exercises will result in answers being
added to your completed assignment.
- Work through the tutorial on bringing up a TurtleBot in Gazebo.
- When the Gazebo window appears, zoom in on the robot and orient the
view of it in such a way that the words on the robot are upside down
and easily legible. Do a screen capture of the Gazebo window at this
point and add this to your completed assignment.
-
Start
rqt_graph
. Do a screen capture of the
rqt_graph
window and add this to your completed
assignment.
- Describe what Gazebo is in a sentence or two and add this to your
assignment. Note that this tutorial does not describe Gazebo. You may
need to search elsewhere for this information. Be sure that this
description is in your own words. Do not simply copy and paste a
description from elsewhere.
- Leave Gazebo and
rqt_graph
running and go on to
the next tutorial.
- Work through the tutorial on exploring the Gazebo world. (Again, note that you should not need to install software on the CSN machines.)
- Explain two differences between controlling the TurtleBot using
turtlebot_teleop
and kobuki_keyop
from the
perspective of the person doing the teleoperation.
- Do a screen capture of
rqt_graph
with each teleop
program running and add these to your completed assignment. Explain
a difference between these two teleop programs from the perspective
of interprocess communication.
- Note that since you loaded the
"TurtleBot world" in the previous tutorial, you won't need to add
objects to the world as described in section 5 of this tutorial.
Instead, you should already have several objects in the world.
Describe these objects and add that description to your completed
assignment.
- Position the robot to face the cylinder and move toward the
cylinder until the curve of the laser on the front of the cylinder is
clearly visible in RViz. Do a screen capture of RViz at this point and
add this to your completed assignment.
- Return to RViz and turn on Image display. Adjust the image
controls until a camera image of the cylinder appears in RViz. Do a
screen capture of RViz at this point and add this to your completed
assignment and explain what you had to adjust in the RViz controls to
obtain a camera image in RViz.
- Adjust the RViz display controls and move the robot until you
obtain a picture of the “rainbow hallway,”
similar to the one shown here.
Capture your own similar image and add this to your completed
assignment. Explain:
- what controls you needed to adjust in RViz to obtain your
rainbow hallway picture,
- what movements you needed to have your simulated TurtleBot
execute to obtain your rainbow hallway picture, and
- what causes the sensor readings to produce each part of the
rainbow hallway image.
- Work through the tutorial on building a map in the TurtleBot simulator.
Note that if you are using the gpel machines and the
associated Docker, there is an error in one of the files—it is
missing part of a path name. Unfortunately, to replace this file
requires superuser priveleges. Fortunately, you have superuser
priveleges in your Docker. Do the following:
- Download my revised
gmapping_demo.launch
file and save it in your ~/Downloads/
directory.
- In a terminal in which you are using your Docker image, use
your superuser priveleges to replace the existing
gmapping_demo.launch
file by typing the following:
sudo cp ~/Downloads/gmapping_demo.launch /opt/ros/indigo/share/turtlebot_gazebo/launch/gmapping_demo.launch
- Use this same terminal when the tutorial tells
you to start map building with
roslaunch turtlebot_gazebo
gmapping_demo.launch
. (If you use a different terminal to
issue this command, it will use the original
gmapping_demo.launch
. Only the terminal in which the
file was replaced will actually use the replacement file.
Note that when you try to launch the TurtleBot playground world, many of the models may fail to load. If any model fails to load, download and install it manually by doing the following:
wget
http://gazebosim.org/models/<model_name>/model.tar.gz
(where <model_name>
is the actual name of the
missing model.
tar xvf model.tar.gz -C ~/.gazebo/models
rm model.tar.gz
- For example, if the dumpster model fails to load, do this:
wget
http://gazebosim.org/models/dumpster/model.tar.gz
tar xvf model.tar.gz -C ~/.gazebo/models
rm model.tar.gz
As you work through the tutorial, don't customize the simulated
TurtleBot—just use it as it is in the launch file.
- When the Gazebo window appears, position it such that you have a
good view of the robot and all of the obstacles in the environment,
then do a screen capture of it and include this in your completed
assignment.
- When the RViz window appears, position it such that you have a good
view of the robot at about the same scale as the robot in the Gazebo
window, then take a screen capture of the RViz window and include that
in your completed assignment.
- Drive the robot around using the keyboard to construct the best map
that you can of the environment. Make sure to drive on all sides of
all objects in the environment. Then take another screen capture of
the RViz window to show the map and include that map in your completed
assignment.
- In your completed assignment, label the screen-captured map from
RViz to show the space that is occupied (occluded), empty (free), and
unknown. Also label each obstacle with its type.
- Be sure to save the map you have created in a form that ROS can use
(using
rosrun map_server map_saver -f <your map
name>
as instructed in the tutorial). Note that you will
include the saved map files (.pgm and .yaml) in your electronic
submission in D2L. Then close out all of your running ROS/Gazebo
processes and move on to the next tutorial.
- Work through the
tutorial on autonomously navigating a known map with a TurtleBot.
However, note that you will be using a simulated TurtleBot for this
assignment. So, rather than bringing up the TurtleBot, you will run
Gazebo as with the previous tutorial.
- When the RViz window appears, position it such that you have a good
view of the robot and the map, similar to screen-captured map you have
already included in your completed assignment for Part 3, above. Take
a new screen capture of the loaded map from RViz and include it in your
completed assignment.
- Explain the various colors that surround the obstacles in the
loaded map in RViz.
- Explain the cluster of green arrows around the robot in RViz.
- Tell the robot its current (starting) position (location and
orientation), following the instructions for how to do so in the
tutorial.
- Give the robot a goal position, following the the instructions for
how to do so in the tutorial. This position should be within the map
you created for the robot but not inside (or extremely close to) one of
the obstacles.
- Take a screen capture of the robot at the goal location and include
this in your completed assignment.
- Give the robot another goal position. This position should be in
the unknown space on your map, outside of the free space. Take a
screen capture of the robot after it executes this command and explain
whether the robot was able to reach this goal.
- Give the robot a third goal position. This position should be
inside one of the mapped obstacles. Take a screen capture of the robot
after it executes this command and explain whether the robot was able
to reach this goal.
What to turn in.
Turn in an electronic copy of this assignment via D2L.