Gazebo Robot Simulation

Gazebo is a 3D robotics simulator that enables rapid testing of algorithms, design of robots in realistic scenarios, and training of AI systems in virtual environments. Originally developed at the University of Southern California by Andrew Howard and Nate Koenig in 2002, Gazebo became the primary simulator for the DARPA Grand Challenge and ROS ecosystem. Key features: high-fidelity physics simulation using ODE, Bullet, Simbody, or DART physics engines for accurate rigid and soft body dynamics, collision detection, and friction modeling. SDF (Simulation Description Format) for describing robots, environments, sensors, and lighting in XML. Realistic sensor simulation including cameras (RGB, depth, thermal), LiDAR (3D and 2D ray casting), IMU, contact sensors, GPS, sonar, and wireless communication models. Plugin architecture for custom controllers, world models, sensors, and visual elements via C++ and Python. Cloud simulation via Gazebo Ignition (now renamed to Gazebo Fortress and Garden) for cloud-native simulation with distributed rendering. GUI for visualizing and interacting with simulation including robot placement, force application, joint control, and camera views. Record and playback of simulation states for reproducible experiments. Integration with ROS for seamless robot control and perception in simulation. World models for indoor, outdoor, and custom environments with terrain, buildings, and dynamic objects. Actor and animation support for pedestrian and vehicle simulation. Parallel simulation for running multiple instances for reinforcement learning training. Used by researchers, educators, and companies for prototyping, testing, and training robot systems before deployment. Open source under Apache 2.0.