Gazebo
Industry-standard open-source robotics simulator with multi-physics and ROS integration
License
Apache-2.0
Overview
Gazebo is an open-source robotics simulator for developing and testing robot software before it runs on real hardware. It provides physics simulation, sensor modeling for cameras, lidars and IMUs, the SDF world description format and a plugin architecture. Integration packages connect it to ROS, so ROS-based software can be tested in simulation.
Gazebo (now Gazebo Sim, formerly Ignition) is the most widely used open-source robotics simulator. It provides high-fidelity physics simulation with ROS integration for developing robot software before deploying to real hardware.
Key features include: - Multiple physics engines (ODE, Bullet, DART, Simbody) - Sensor simulation (cameras, lidars, IMUs, etc.) - GPU-accelerated rendering - SDF world description format - ROS 1/2 integration packages - Plugin architecture for customization - Large model database - Distributed simulation support
Gazebo is used by robotics teams worldwide for algorithm development, testing, and validation. It's essential for any ROS-based robotics project.
Why it matters
Gazebo has been the default robotics simulator for over a decade. Its deep ROS integration means that nearly every ROS-based robot has been tested in Gazebo first, making it the most battle-tested simulation platform in the robotics community.
Problem it solves
Developing and testing robot software on real hardware is expensive, slow, and risky. Gazebo provides a complete simulation environment where ROS-based software can be tested identically to how it runs on real robots.
Key takeaways
- Industry-standard robotics simulator with deep ROS integration
- Accurate physics with Bullet and DART backends
- Comprehensive sensor simulation including cameras, LIDAR, and IMU
- Large model library with community-contributed robots and environments
- Active development as Gazebo Sim (formerly Ignition Gazebo)
Who it is for
- ROS developers testing robot software
- Robotics companies validating algorithms before hardware deployment
- Researchers needing full robot simulation with sensors
- Educators teaching robotics with simulation
Limitations
- Performance not competitive with GPU-accelerated simulators for RL training
- Transition from Classic to Gazebo Sim created ecosystem fragmentation
- Physics accuracy varies between different backends
- Rendering quality functional but not photorealistic