Internship Progress Week 2 (July 9 ~ July 15)
This week, my main task was to install, configure, and run the Gazebo simulator using three distinct methods and verify their working with a custom robot simulation.
Method 1: Local Binary Package Installation (ROS 2 Humble)
The first method was installing Gazebo directly on the host system using official ROS 2 Humble binary packages via apt.
To verify its working, I launched Gazebo and Echoed the /lidar sensor topic inside the host terminal. The output shows real-time lidar range coordinates successfully communicating over the ROS 2 network.
1
2
3
4
sudo apt update
sudo apt install -y ros-humble-ros-gz-sim ros-humble-ros-gz-interfaces ros-humble-ros-gz-bridge
# Verify sensor working
gz topic -e -t /lidar
Method 2: Local Source Build (colcon workspace)
The second method involved compiling Gazebo directly from source. I created a dedicated colcon workspace (gazebo_source_ws), checked out the source repositories, and compiled them manually.
To verify the working of the source build, I sourced the local overlay and launched the Gazebo GUI, which successfully initialized the simulation environment.
1
2
3
cd ~/gazebo_source_ws
source install/setup.zsh
ruby $(which gz) sim -g
Method 3: Ubuntu VNC Docker Container (tiryoh/ros2-desktop-vnc)
The third method used a Dockerized Ubuntu container pre-configured with ROS 2 Humble and a VNC server (tiryoh/ros2-desktop-vnc:humble). This allows running Gazebo in an isolated containerized Ubuntu environment while visualizing the graphical interface through a web browser on localhost:6080.
To verify its working, I accessed the container’s Ubuntu desktop, sourced ROS 2, launched Gazebo, and spawned the custom robot inside the simulated arena.
1
2
3
# Inside Docker Ubuntu Container
source /opt/ros/humble/setup.bash
ign gazebo /home/ubuntu/laser_world.sdf
