Mobile Robot Simulation

This is a tutorial on robot localization with the tools of ROS and RViz. All the code which is required is included in the catkin_ws folder.
This demo was tested on Ubuntu 20.04 using ROS Noetic Ninjemys.
We will test the following examples:

• Controlling the Robot in Simulation
• Viewing Sensor Data
• Creating a Map // World

Getting Started

We will run the Rviz simulator of virtual robot Turtlebot3 in version of ROS Noetic Ninjemys.
We will test SLAM (Simultaneous localization and mapping) and autonomous navigation.

Prerequisites

We assume you have ROS Noetic installed using Ubuntu 20.04. Check which version of ROS you have by running the command:

$ ls /opt/ros

Installation

We need to install all of the dependent packages of TurtleBot3 simulator.

We first make a ROS workspace that will contain all the code we will working on.
In a new terminal, enter the following commands (one right after the other):

$ mkdir -p ~/catkin_ws/src

Clone the source code (You need to have git installed):

$ cd ~/catkin_ws/src/
$ git clone https://github.com/ROBOTIS-GIT/turtlebot3_msgs.git
$ git clone https://github.com/ROBOTIS-GIT/turtlebot3.git
$ git clone https://github.com/ROBOTIS-GIT/turtlebot3_simulations.git

Then, we need to build the files:

$ cd ~/catkin_ws/
$ catkin_make

Next, source the workspace:

source ~/catkin_ws/devel/setup.bash

Note: TurtleBot3 has three models (Burger, Waffle, and Waffle Pi). We set the model Burger before we launch TurtleBot3.
This command open the bashrc file to add this setting:

gedit ~/.bashrc

And we write this line at the bottom of the file: export TURTLEBOT3_MODEL=burger
Save and close the file, then reload .bashrc to log back in:

source ~/.bashrc

Simulating TurtleBot3 with RViz

RViz is a physics engine in which we will run our simulation.
The command roslaunch enables us to launch a program. ROS applications are organized in a system of packages, each with its own launch file. When we call roslaunch, we need to specify the desired package and launch file. We can build each ROS package alone. It is the smallest functional unit in the workspace. Note that we need to be inside a package for every ROS program we write.

The ROS package should at least contain these elements:

• src folder: Source code (C++, Python)
• CMakeLists.txt: CMake rules for compilation
• package.xml: Package information and dependencies

We launch the simulation using the following command:

roslaunch turtlebot3_fake turtlebot3_fake.launch

If we want to move the robot around his environment, we need another launch file. We type this command in a new terminal:

roslaunch turtlebot3_teleop turtlebot3_teleop_key.launch

In this terminal we click on the mentioned keys to control the movement of TurtleBot3:

The Rviz simulator should open with the following screen:

SLAM and Autonomous Navigation with TurtleBot3

SLAM concerns the problem of a robot building or updating a map of an unknown environment while simultaneously keeping track of its location in that environment.

We choose an environment with obstacle avoidance for testing SLAM and navigation algorithms. The goal is to make TurtleBot3 autonomously navigate around a room and avoid colliding into objects.
As a first step, we can access and inspect the sensor data. We can inspect the scan data from rostopic using a new terminal:

$ rostopic info /scan

We can check its publishing messages using:

$ rostopic echo /scan

The robot is being simulated in an open world, with no obstacles in sight, so we manually add an obstacle. We check the laser scanner topic again as we should notice that the output has changed.
We can open RViz to visualize the LaserScan topic while TurtleBot3 is moving around in the world. In a new terminal, type:

roslaunch turtlebot3_gazebo turtlebot3_gazebo_rviz.launch

So we can see the point cloud:

RViz is able to display other data. For instance, we can view the vision data by selecting the camera topic. The output of this would look like:

ROS comes with the SLAM module. Install it in a new terminal:

sudo apt install ros-noetic-slam-gmapping

We download the world we seek to map by calling this command:

roslaunch turtlebot3_gazebo turtlebot3_world.launch

Then, start SLAM in a new terminal:

roslaunch turtlebot3_slam turtlebot3_slam.launch slam_methods:=gmapping

Finally, launch autonomous navigation in a last terminal tab:

roslaunch turtlebot3_gazebo turtlebot3_simulation.launch

We can watch the robot create a map of the environment as it autonomously moves from place to place!

And that's it!
Keep building ;)