Let's dive into Gazebo – a leader in robot simulations.
Gazebo is an open-source 3D robot simulator, with ROS (Robot Operating System) serving as the interface for the robot. It's a prominent tool on which a myriad of users and developers all over the world rely.
A well-designed simulator allows you to quickly design robots, test algorithms, run regression tests, and train AI systems with the use of authentic scenarios. Gazebo makes it possible to accurately and efficiently simulate robots populations in complex environments, both indoor and outdoor ones.
Gazebo is a potent environment of 3D simulation intended for autonomous robots. It's fit especially for testing computer vision and object avoidance. The tool allows you to create a 3D scenario with robots, obstacles, and plenty of other objects on your computer. It provides high-quality graphics as well as convenient graphical and programmatic interfaces. What's more, Gazebo uses a physics engine for gravity, illumination, inertia, among others. The simulator enables you to assess and test your robot in challenging scenarios not putting the actual robot at any risk.
When it comes to the Gazebo origin, its development started at the University of Southern California in 2002. The need to simulate robots in an outdoor environment under a variety of conditions was the very reason the idea of a high-fidelity simulator arose from.
At first, the purpose of Gazebo was to be used for evaluating algorithms for robots. In many cases, it's essential to test the robot applications like battery life, error handling, localization, navigation, and grabbing. But once the need for a multi-robot simulator emerged, Gazebo was improved to serve as a tool it is today.
The name Gazebo was chosen as the structure most closely resembling an outdoor scene. Although the majority of Gazebo users simulate indoor environments, the name remained.
As since 2020 we've partnered with the European Rover Challenge to co-organize the remote edition (see more in the article), we're constantly developing more and more in a dedicated simulation.
The simulation we had created for the event consisted of a model of the Mars yard, the software, and naturally, Leo Rover along with the attached gear, including the cameras: Navcam – ZED 2 and Hazcam – RPi. We had a drone take photos and 3D scan the area which was the key in creating a three-dimensional model of the terrain.
One of the features that distinguished Leo Rover in the Gazebo simulation for the ERC was that the robot functioned very much like the real one. The contestants could write their own software for the simulation and use it as if they were working with the actual rover on the real terrain.
Another thing worth pointing out regarding Leo Rover in the simulation was a special plugin created by Błażej from our team. It allowed the rover to turn on the spot. The plugin worked on the force of the wishbones keeping them in the correct position, just like in the real robot.
And here is a video demonstrating the work of Space Concordia - Robotics Division in the Gazebo simulation used for developing their software for the Simulation Task for the ERC 2021:
Also, check out the following youtube channel to see Leo Rover roaming across a simulated terrain.
So, we’ve shed some light on Gazebo. As seen above, it’s a tool of paramount importance for users all around the world. Maybe you’re using it for your simulations as well or are willing to try? If so, see our tutorial on the Gazebo simulation of the Navigation Task for the ERC: