google-deepmind/mujoco_menagerie

A collection of high-quality models for the MuJoCo physics engine, curated by Google DeepMind.

What it solves

MuJoCo Menagerie provides a curated library of high-quality, well-designed robot models for the MuJoCo physics engine. It solves the problem of "bad" models that do not behave as expected, ensuring that researchers and developers have access to reliable simulations that work immediately upon deployment.

How it works

The project organizes robot models into a standardized structure, providing MJCF XML definitions, 3D meshes (STL or OBJ), and scene files for easy loading. Some models include MJX-compatible versions for accelerated simulation. Users can integrate these models into their projects either by cloning the repository or using the robot-descriptions Python package for direct loading.

Who it’s for

It is designed for robotics researchers, AI developers, and engineers who use the MuJoCo physics engine to simulate robots, train agents, and test control algorithms.

Highlights

  • Extensive Library: Includes a wide variety of robots including humanoids, quadrupeds, bipeds, drones, and robotic arms.
  • Quality Grading: Implements a grading system (A+ to C) to communicate the faithfulness of a model to its real-world counterpart.
  • MJX Support: Provides specialized variants for models compatible with MJX for high-performance simulation.
  • Standardized Format: Each model includes a-la-carte assets, XML definitions, and scene files for consistent implementation.

Related

  • Project
    ARISE-Initiative/robosuite

    A MuJoCo-powered simulation framework and benchmark suite for robot learning, providing modular tools to create environments and test AI control algorithms.

  • Project
    google-deepmind/mujoco

    A high-performance physics engine for the fast and accurate simulation of articulated structures, widely used in robotics and machine learning research.

  • Project
    google-deepmind/mujoco_playground

    A GPU-accelerated suite of robot learning environments built with MuJoCo MJX, designed for efficient reinforcement learning and sim-to-real research.

  • Project
    google-deepmind/mujoco_mpc

    A software framework for real-time predictive control using MuJoCo, enabling the authoring and solving of complex robotics tasks through various planning methods.

  • Project
    vikashplus/robohive

    A unified framework for robot learning that provides a collection of MuJoCo-simulated environments and tasks exposed via the OpenAI-Gym API.