Gazebo (robotics simulator)

Gazebo (robotics simulator)
Name	Gazebo
Developer	Open Robotics
Released	2004
Programming language	C++, Python
Operating system	Linux, Windows, macOS
License	Apache License 2.0

Gazebo (robotics simulator) is an open-source three-dimensional simulator for testing and developing robotic systems in virtual environments. It provides realistic dynamics, sensor emulation, and model libraries to support research and development by institutions such as Stanford University, Massachusetts Institute of Technology, Carnegie Mellon University, and companies including Google, Amazon (company), and Toyota Motor Corporation. Gazebo is widely used alongside frameworks and projects from Open Source Robotics Foundation, ROS (Robot Operating System), PX4, and Autoware.

Overview

Gazebo offers physics-based simulation, graphical rendering, and plugin infrastructure to model robots, environments, and interactions with fidelity comparable to testbeds at NASA, DARPA, and European Space Agency. It supports multi-robot scenarios, complex terrains, and integration with middleware such as ROS 1, ROS 2, DDS, and YARP. The platform enables reproducible experiments used by research groups at University of Oxford, ETH Zurich, University of Tokyo, and industrial labs at ABB, Siemens, and Bosch.

History and Development

Development began in the early 2000s by researchers associated with Willow Garage and later continued under the Open Source Robotics Foundation. The project evolved alongside milestones from DARPA Robotics Challenge and collaborations with European Space Agency for planetary robotics work. Major contributors and maintainers have included teams from SRI International, NASA Jet Propulsion Laboratory, University of Southern California, and corporate sponsors such as Intel Corporation and NVIDIA. The simulator has undergone major rewrites to improve modularity, connect with ROS 2, and support modern graphics and compute stacks championed by Khronos Group and OpenGL evolutions.

Architecture and Features

Gazebo's architecture separates world description, physics, rendering, and plugins; it uses model description formats compatible with URDF and SDF (Simulation Description Format). Core components include a server for physics stepping, client for rendering, and plugin APIs for custom controllers used by teams at Toyota Research Institute and labs at Imperial College London. Features encompass scene graph rendering, texture and lighting managed through drivers from NVIDIA Corporation and AMD, and networked simulation influenced by standards from Object Management Group. The system supports headless execution for continuous integration setups used by projects at GitHub, Travis CI, and Jenkins.

Supported Sensors and Actuators

Gazebo simulates sensors such as cameras, depth sensors, LiDARs, IMUs, and GPS units used by platforms from DJI, Clearpath Robotics, and Boston Dynamics. Actuator models include wheel joints, articulated arms, grippers, and thrusters used in collaborations with Blue Origin and SpaceX prototypes; these are commonly controlled via interfaces standardized in ROS Control and flight stacks like PX4 Autopilot. Sensor plugins emulate noise models and calibration behaviors informed by specifications from Velodyne, SICK AG, and FLIR Systems.

Physics Engines and Performance

The simulator integrates physics engines including ODE (software), Bullet (software), DART (software), and MuJoCo to handle rigid-body dynamics, contact resolution, and soft-body interactions relevant to work at MIT CSAIL and Stanford Artificial Intelligence Laboratory. Performance tuning leverages multi-threading on processors by Intel and vectorization on GPUs from NVIDIA with CUDA optimizations; benchmarking is common in labs at ETH Zurich and industry testing at Ford Motor Company and General Motors.

Integrations and Ecosystem

Gazebo is tightly integrated with ecosystems like ROS 1, ROS 2, Ignition (robotics), and mapping stacks such as GMapping and Cartographer (software). It supports payloads, models, and world files contributed through community repositories on GitHub and distribution channels used by Ubuntu and Debian. Interfacing with autonomy stacks from Autoware and perception frameworks like OpenCV and PCL (Point Cloud Library) is common in field trials coordinated with partners such as Waymo and Uber ATG.

Use Cases and Applications

Gazebo is used for autonomous vehicle simulation by research teams at California Institute of Technology and industry groups at NVIDIA DRIVE, robotic manipulation research at MIT, aerial robotics testing for companies like Skydio, and space robotics planning in cooperation with NASA. It facilitates curriculum development at institutions including Georgia Institute of Technology, University of California, Berkeley, and Princeton University and supports competitions such as RoboCup, DARPA Subterranean Challenge, and university-level hackathons sponsored by IEEE.

Community and Licensing

The project is maintained under the Apache License 2.0 and governed by contributor agreements and stewardship from organizations like Open Robotics and collaborating institutions including Willow Garage alumni and corporate engineering teams from Google, Amazon, and Toyota. A vibrant community engages through mailing lists, forums, and code hosting on GitHub with reproducible workflows using continuous integration services from GitLab and cloud providers such as Amazon Web Services, Google Cloud Platform, and Microsoft Azure.

Category:Simulation software