Robot Models

A robot model is a structured digital description of a robot’s kinematic and dynamic properties. It defines the robot’s physical structure (links and joints), geometry, mass, and inertial parameters. Robot models are essential for tasks such as state estimation, simulation, collision checking, inverse kinematics, and inverse dynamics.

These models are used by physics engines and robotics frameworks like ROS, Gazebo, and MuJoCo to simulate motion, compute control strategies, and visualize robot behavior.

Common formats include:

• URDF (Unified Robot Description Format): Widely used in ROS, URDF describes a robot’s links, joints, visuals, collisions, and basic physical properties.

• SDF (Simulation Description Format): Used in Gazebo, SDF extends URDF with support for sensors, plugins, environmental elements, and more advanced simulation features.

• MJCF (MuJoCo XML Format): The native format for MuJoCo, MJCF defines detailed models of robots and environments, including joints, actuators, geometries, materials, and simulation settings. It is optimized for physics-based control and research.