Week 1

Week 1: ROS and Linux Basics

You should be able to complete this week's work within the allotted 2-hour lab session.

Note: There are two intensional errors planted in this week's lab material! Can you find them and work out how to fix them?!

Introduction

In this first week you will learn the basics of ROS and become familiar with some of the key tools and principles of this framework, which will allow you to program robots and work with ROS applications effectively. For the most part, you will interact with ROS using the Linux command line and so you will also become familiar with some key Linux command line tools that will help you. Finally, you will learn how to programme some basic ROS Nodes using Python.

Intended Learning Outcomes

By the end of this week's lab session you will be able to:

1. Control a TurtleBot3 Robot, in simulation, using ROS
2. Launch ROS applications using roslaunch and rosrun
3. Interrogate running ROS applications using key ROS command line tools
4. Build a ROS package containing multiple nodes, and program these nodes (in Python) to communicate with one another using ROS Communication Methods
5. Navigate a Linux filesystem and learn how to do various filesystem operations from within a Linux Terminal

Packages

ROS applications are organised into packages. ROS packages are collections of scripts, parameters and configurations relating to some common robot functionality, and ROS uses packages as a way to organise all the programs running on a robot. The package system is a fundamental concept of ROS, and all ROS programs are organised in this way.

1. If you haven't done so already, launch your WSL-ROS environment by running the WSL-ROS shortcut in the Windows Start Menu. This should open up a terminal application and an Ubuntu terminal instance. We'll refer to this terminal instance as TERMINAL 1

2. In the terminal enter the following command to launch a simulation of a TurtleBot3 Waffle in an empty world:

    [TERMINAL 1] $ roslaunch turtlebot3_gazebo turtlebot3_empty_world.launch
   

   Note: A $ symbol indicates a command that should be entered directly into a Linux Terminal. Enter (or copy and paste) all the text after the $ into the terminal, excluding the $ itself!

3. A Gazebo simulation window should open and within this you should see a TurtleBot3 Waffle (similar to the real robots that you will be working with later this semester) in empty space:

   

Exercise 1: Making the robot move

1. With your Gazebo Simulation up and running, return to the terminal application and open up a new Ubuntu terminal instance (TERMINAL 2) by pressing the New Tab button:

   

   (or, alternatively, press the Ctrl+Shift+T keyboard shortcut). In the new terminal instance enter the following command:

    [TERMINAL 2] $ roslaunch turtlebot3_teleop turtlebot3_teleop_key.launch
   

2. Follow the instructions provided in the terminal to drive the robot around in its simulated environment

Summary:

You have so far launched two separate ROS applications using the roslaunch command. roslaunch is one way to launch ROS programs. As you should have observed from the above examples, we use this command in the following way:

roslaunch [package name] [launch file]

The command takes two parameters as inputs: [package name] is the name of the ROS package containing the functionality that we want to execute and [launch file] is a file within that package that tells ROS exactly what functionality within the package we want to launch.

Exercise 2: Exploring a Package

roscd is a ROS command that allows us to navigate to the directory of any ROS package installed on our system, without us needing to know the path to the package beforehand.

1. Open up a new terminal instance (TERMINAL 3) and use the roscd command to navigate to the turtlebot3_teleop package directory on our Linux filesystem:

    [TERMINAL 3] $ roscd turtlebot3_teleop
   

   The terminal prompt should have changed to illustrate where on the filesystem the roscd command has just taken you:

   

2. pwd is a Linux command which tells us where the terminal is currently located in the filesystem (i.e. Print Working Directory). Use this to confirm what the terminal prompt is telling you:

    [TERMINAL 3] $ pwd
   

   We now know where the turtlebot3_teleop package is located on our machine, and we can then use more Linux commands to explore this further:

3. ls is a Linux command which lists the contents of the current directory. Use this to list the contents of the turtlebot3_teleop package directory.

    [TERMINAL 3] $ ls
   

   This will simply list the items in the current directory, but we can use the -F option to find out a little more:

    [TERMINAL 3] $ ls -F
   

   You will notice that the output has now changed slightly: items followed by a / are folders (aka "directories") and items without the / are files (files will often have a file extension too). How many items are there in the turtlebot3_teleop package directory? How many of these are directories and how many are files?

   Launch files for a package are typically located in a launch folder within the package directory. You should have noticed a launch folder in the output of the ls command above.

4. cd is a Linux command that allows us to Change the Directory that the terminal is currently located in. Use this to navigate to the turtlebot3_teleop package launch folder and then use ls again to see what's in there. In this folder you should see the turtlebot3_teleop_key.launch file that we executed with the roslaunch command in Exercise 1. We will now have a look at the contents of this file...

5. cat is a Linux command that we can use to display the contents of a file in the terminal. Use this to display the contents of the turtlebot3_teleop_key.launch file.

    [TERMINAL 3] $ cat turtlebot3_teleop_key.launch
   

Summary:

From the output of cat in the step above you should have noticed that the contents of a launch file are contained within a <launch> tag:

<launch>
... 
</launch>

Within that, we also have (amongst other things) a tag which tells ROS exactly what scripts or executables to launch and how to launch them:

<node pkg="turtlebot3_teleop" type="turtlebot3_teleop_key" name="turtlebot3_teleop_keyboard"  output="screen">
</node>

The attributes here have the following meaning:

• pkg: The name of the ROS package containing the functionality that we want to launch
• type: The name of the script (or ROS Node) that we want to execute within that package (when we create our own Python scripts we need to make sure that we include the .py file extension here too)
• name: A descriptive name that we want to give to the ROS node
• output: The place where any output from the node will be printed (either screen where the output will be printed to our terminal window, or log where the output will be printed to a log file)

ROS Nodes are executable programs that perform specific robot tasks and operations, such as remote (or "teleoperational") control, as we have seen in the above example.

The turtlebot3_teleop package that we have just interrogated here is a package that has been installed on our Linux machine, so the package contents that we have just viewed are fairly minimal, and the Nodes (the source code files) are actually located elsewhere on the system. In fact, the turtlebot3_teleop_key node is actually located in the /opt/ros/melodic/lib/ directory and you could use the Linux/ROS command line tools that you have learnt about above to interrogate this file, if you wanted to....

When we create our own packages and nodes these will be located in the package directory as well (unlike the installed package above). We will learn more about this in a later exercise.

A ROS Robot may have hundreds of individual nodes running simultaneously to carry out all its necessary operations and actions. Each node runs independently, but uses ROS communication methods to communicate and share data with other nodes on the system.

Publishers and Subscribers: A ROS Communication Method

You can use the ROS command rosnode to view all the nodes that are currently active on a ROS system.

You should currently have three terminal windows active: the first in which you launched the Gazebo simulation (TERMINAL 1), the second with your turtlebot3_teleop node active (TERMINAL 2), and the third where you explored the contents of the turtlebot3_teleop package directory (TERMINAL 3). TERMINAL 3 should now be idle.

Exercise 3: Visualising the ROS Network

1. In TERMINAL 3 enter $ cd ~ to go back to your home directory (remember that ~ is an alias for this)

2. Use the following command to have a look at what nodes are currently active:

    [TERMINAL 3] $ rosnode list
   

   Only a handful of nodes should be listed:

    /gazebo
    /gazebo_gui
    /rosout
    /turtlebot3_teleop_keyboard
   

3. We can visualise the connections between the active nodes by using the rqt_graph node within the rqt_graph package. We can use rosrun to launch this node directly (you might get some error messages, but don't worry about them):

    [TERMINAL 3] $ rosrun rqt_graph rqt_graph
   

   A new window should then open, displaying something similar to the following (hover over the diagram to enable colour highlighting):

   

   This tool shows us that the /turtlebot3_teleop_keyboard and /gazebo nodes are communicating with one another. The direction of the arrow tells us that the /turtlebot3_teleop_keyboard node is a Publisher and the /gazebo node is a Subscriber. The two nodes communicate via a ROS Topic, in this case the /cmd_vel topic, and on this topic the /turtlebot3_teleop_keyboard node publishes messages.

Exercise 4: Exploring ROS Topics and Messages

We can find out more about the /cmd_vel topic by using the rostopic ROS command.

1. In a new terminal instance (TERMINAL 4) type the following:

    [TERMINAL 4] $ rostopic info /cmd_vel
   

   This should provide an output similar to the following:

    Type: geometry_msgs/Twist
   
    Publishers:
     * /turtlebot3_teleop_keyboard (http://localhost:#####/)
   
    Subscribers:
     * /gazebo (http://localhost:#####/)
   

   This confirms what we discovered earlier about the publisher(s) and subscriber(s) to the /cmd_vel topic. In addition, this also tells us the topic type, or the type of message that is being published on this topic.

2. We can use the rosmsg ROS command to provide further information about this message, or any other message that we may be interested in:

    [TERMINAL 4] $ rosmsg show geometry_msgs/Twist
   

   From this, we should obtain the following:

    geometry_msgs/Vector3 linear
      float64 x
      float64 y
      float64 z
    geometry_msgs/Vector3 angular
      float64 x
      float64 y
      float64 z
   

   We'll learn more about what this means next week.

3. To finish, shut down any active terminal processes by entering Ctrl+C in any that still have processes running (Terminals 1, 2 and 3). The associated Gazebo and rqt_graph windows should close as a result of this too.

Exercise 5: Creating a Package

In the following exercises we will create some simple publisher and subscriber nodes in Python and send messages between them. As we learnt earlier though, ROS applications should be contained within packages and so we need to create a package in order to start creating our own ROS nodes.

ROS provides a tool to create a new ROS package and ensure that all the essential elements are present: catkin_create_pkg.

It is important to work in a specific filesystem location when we create and work on our own ROS packages, so that ROS can access and build everything appropriately. These spaces are called "Catkin Workspaces" and one has been created for you already, called catkin_ws.

1. We can navigate to the catkin_ws folder by using the Linux cd command. In TERMINAL 1 enter the following:

    [TERMINAL 1] $ cd ~/catkin_ws/
   

2. Inside the catkin workspace there is a directory called src (use the ls command to confirm this). All new packages need to be located in the src folder, so we need to be here when we use the catkin_create_pkg tool to create a new package. So, use the cd command again to navigate to the catkin_ws/src folder:

    [TERMINAL 1] $ cd src
   

3. Now, use the catkin_create_pkg script to create a new package called pubsub_example which will have std_msgs and rospy as dependencies:

    [TERMINAL 1] $ catkin_create_pkg pubsub_example std_msgs rospy
   

   What did the catkin_create_pkg tool just do? (Hint: there are four things and it will have told you about them!)

4. Navigate into this new package directory and use ls to list the content that has been created by the catkin_create_pkg tool.

Catkin packages are typically organised in the following way, and have a few essential features that must be present in order for the package to be valid. These items are highlighted with [*]:

package_folder/      -- All packages must be self-contained within their own root folder [*]
  |-launch/          -- A folder for launch files
  |-src/             -- A folder for source files (python scripts etc)
  |-CMakeLists.txt   -- Rules for compiling the package [*]
  `-package.xml      -- Information about the package [*]

You will have noticed that the catkin_create_pkg tool made sure that the essential features of a Catkin Package were created when we asked it to build the pubsub_example package above.

Exercise 6: Creating a Publisher node

1. Within the pubsub_example package directory, navigate to the src folder using the cd command

2. touch is a Linux command that we can use to create an empty file. Use this to create an empty file called publisher.py, which we will add content to shortly:

    [TERMINAL 1] $ touch publisher.py
   

3. Because we want to execute this script eventually, we will need to give the file execute permissions, which may not have been set by default when the file was created in the previous step. To do this, we can use the Linux chmod command in the following way: chmod +x [name of the python script]. First though have a look at the file as it is using ls again, but this time with an additional option:

    [TERMINAL 1] $ ls -lF  
   

   Which should provide the following output:

    -rw-r--r-- 1 student student 0 Jan 01 12:00 publisher.py
   

   The first bit of the output here tells us the file permissions: -rw-r--r--. This tells us who has permission to do what with this file and currently, the first bit: -rw-, tells us that we have permission to Read or Write to it.

   Now run the chmod command:

    [TERMINAL 1] $ chmod +x publisher.py
   

   And then run the ls -lF command again to see what has changed:

    [TERMINAL 1] $ ls -lF
    
    -rwxr-xr-x 1 student student 0 Jan 01 12:00 publisher.py*
   

   We can now see that we have granted everyone (including ourselves) permission to eXecute the file too. Job done.

4. We can now open this file in a text editor. In this case we will use Atom. Launch this app now from the command line:

    [TERMINAL 1] $ atm
   

5. Then we need to open the publisher.py file that we have just created. You can do this by either navigating to it using the file tree to the left-hand-side of the Atom window, or by entering the following in TERMINAL 1:

    [TERMINAL 1] $ atm publisher.py
   

6. Once opened, copy the code provided here into the empty file and save it. Note: It is important that you understand how this code works, so you should have a look at the explainer, provided below the code.

7. We can now run this node using the ROS command rosrun. However, we closed everything down earlier on, so the ROS Master is no longer active. First then, we need to re-launch it manually using roscore:

    [TERMINAL 1] $ roscore
   

8. Then, in TERMINAL 2, use rosrun to execute the publisher.py script that you have just created by providing the relevant information to the rosrun command as follows: rosrun [package name] [script name] If you see a message in the terminal similar to the following then the node has been launched successfully:

    publisher node is active...
   

   We can further verify that our publisher node is running using a number of different tools. Try the following in TERMINAL 3:

9. $ rosnode list: This will provide a list of all the nodes that are currently active on the system. Verify that the /publisher_node is visible in this list.

10. $ rostopic list: This will provide a list of the topics that are currently being used by nodes on the system. Verify that the /chatter topic is present within this list.

Using the rostopic command

So far we have used the rostopic ROS command with two additional arguments:

• list to provide us with a list of all the topics that are active on our ROS system, and
• info to provide us with information on a particular topic of interest.

We can use the autocomplete functionality of the Linux terminal to provide us with a list of all the available options that we can use with the rostopic command. To do this you can type rostopic followed by a space and then press the Tab key twice:

rostopic[SPACE][TAB][TAB]

You should then be presented with a list of the available arguments for the rostopic command:

• rostopic hz [topic name] provides information on the frequency (in Hz) at which messages are being published to a topic:

    [TERMINAL 3] $ rostopic hz /chatter
  

  This should tell us that our publisher node is publishing messages to the /chatter topic at (or close to) 10 Hz, which is exactly what we ask for in the publisher.py file (in the __init__ part of our Publisher class). Press Ctrl+C to stop this command.

• rostopic echo [topic name] shows the messages being published to a topic:

    [TERMINAL 3] $ rostopic echo /chatter
  

  This will provide a live stream of the messages that our publisher.py node is publishing to the /chatter topic. Press Ctrl+C to stop this.

• We can see some additional options for this command by viewing the help documentation:

    [TERMINAL 3] $ rostopic echo -h
  

  From here, for instance, we can learn that if we just wanted the echo command to display a set number of messages from the /chatter topic we could use the -n option. To display the most recent two message only for example:

    [TERMINAL 3] $ rostopic echo /chatter -n2
  

Exercise 7: Creating a Subscriber Node

You will now create another node to subscribe to the /chatter topic and access the information that our publisher node is outputting.

1. In TERMINAL 3 use the filesystem commands that were introduced earlier (cd, ls and roscd) to navigate to the src folder of the pubsub_example package that we created earlier.

2. Use the same procedure as before to create a new empty Python file called subscriber.py and make it executable.

3. Then, open the newly created subscriber.py file in Atom, paste in the code provided here and save it. Once again, it's important that you understand how this code works, so we have provided an explainer for this too.

4. Use rosrun (remember: rosrun [package name] [script name]) to run your newly created subscriber.py node. If your publisher and subscriber nodes are working correctly you should see an output like this:

   

5. As before, we can find out what nodes are running on our system by using the $ rosnode list command. Open a new terminal window (TERMINAL 4), run this and see if you can identify the nodes that you have just launched.

6. Finally, close down your publisher and subscriber nodes and the ROS Master by entering Ctrl+C in Terminals 1, 2 and 3.

Launch Files

At the beginning of this lab session we launched our Gazebo Simulation and the turtlebot3_teleop_keyboard node using launch files and the roslaunch command. This provides a means to launch multiple ROS nodes simultaneously, and we will demonstrate this by building a launch file for the publisher and subscriber nodes that we have just created.

Exercise 8: Creating a launch file

1. In TERMINAL 1, use roscd to navigate to the pubsub_example package directory.

2. Use the Linux mkdir command to make a new directory in the package root folder called launch:

    [TERMINAL 1] $ mkdir launch
   

3. Use the cd command to enter the launch folder that you have just created, then use the touch command (in the same way as before) to create a new, empty file called pub_sub.launch.

4. Open this launch file in Atom and enter the following text:

    <launch>
      <node pkg="pubsub_example" type="publisher" name="pub_node"  output="screen">
      </node>
    </launch>
   

5. Use roslaunch to launch this file and test it out as it is (remember: roslaunch [package name] [launch file]). If everything looks OK then carry on to the next step.

6. The launch file that you have just created should launch the publisher.py node, but not the subscriber.py node. Add another <node> tag to launch the subscriber node as well.

7. The publisher and subscriber nodes and the ROS Master can now all be launched with the roslaunch command and the pub_sub.launch file that you have now created.

8. Launch this in TERMINAL 1 and then use $ rosnode list in TERMINAL 2 to check that it all works correctly.

Summary:

• roslaunch can be used to launch multiple nodes on a robot from one single command.
• It will also automatically launch the ROS Master (equivalent to running the roscore command manually) if it isn't already running. Did you notice that we didn't have to do this manually in this exercise, but we did when we launched our nodes individually using rosrun?
• In the rospy.init(...) functions of our publisher.py and subscriber.py Python scripts, we defined a node name and set anonymous=True. As a result, when we launched our nodes manually using rosrun, the names we defined were honoured, but were appended with a unique combination of numbers.
• When we launched our nodes using roslaunch however, the node names were set according to what we had defined in the name field of the <node> tag within the launch file, and anything specified within the rospy.init(...) function within our Python scripts was overwritten as a result.

Wrapping Up

In this session we've learnt about some of the key principles of ROS such as Packages, Launch files, Nodes and the Publisher-Subscriber Communication Method using Topics and Messages.

We have learnt how to use some key ROS commands:

• roslaunch: to launch multiple ROS Nodes via launch files
• roscd: to navigate to installed ROS packages using a package name alone
• rosnode: to display information about active ROS Nodes
• rosrun: to run executables within a ROS package
• rostopic: to display information about active ROS topics
• rosmsg: to display information about all ROS messages that are available to use in a ROS application
• roscore: to launch the ROS Master: The baseline nodes and programs that are required for ROS to function

In addition to this we also learnt how to use catkin_create_pkg, which is a helper script for creating ROS package templates.

We have also learnt how to work in the Linux Terminal and navigate a Linux filesystem using key commands such as:

• pwd: prints the path of the current working directory to show you which directory you're currently located in
• ls: lists the files in the current directory
• cd: change directory to move around the file system
• mkdir: make a new directory (mkdir [new_folder])
• cat: show the contents of a file
• chmod: modify file permissions (i.e. to add execute permissions to a file for all users: chmod +x [file])
• touch: create a file without any content

Finally, we have learnt how to create basic ROS nodes in Python to both publish and subscribe to ROS topics using standard ROS messages.

Saving your work

Remember, the work you have done in this WSL-ROS environment today will not be preserved for future sessions or across different University machines automatically! To save the work you have done here today you should now run the following script in any idle WSL-ROS Terminal Instance:

$ rosbackup.sh

This will export your home directory to your University U: Drive, allowing you to restore it at the start of next week's session.

Navigating This Wiki:
← Getting Started | Week 2: Odometry and Basic Navigation →