The Barkour Robots are a series of agile quadruped robots developed by Google DeepMind to support various research projects.
This repository contains the design assets (CAD, PCBA designs, assembly instructions...) and core software (firmware and low level code) of the Barkour robots.
Detailed documentation is available in the docs/ folder:
- Robot getting started/basic usage guide
- Software:
- Hardware:
- Robot assembly guides:
Mechanical CAD files and simulation environments are available at the following links:
- CAD model
- Simulation:
We release 2 versions of the Barkour Robot. One version has a fixed head, the other has an actuated head with a mouth (gripper). You can find the CAD models here for both versions.
Please take a look at the getting started guide.
The simulation models were previously released as part of the MuJoCo Menagerie. You can find the MuJoCo model here.
We support MJX for fast training on GPUs and TPUs. Please take a look at the MJX tutorial to train a policy in minutes.
We are including the schematics and PCB layout files for the custom electronics
on the Barkour robots. Take a look in the hardware/electronics folder.
You can find drawings of all custom cable assemblies in the
hardware/cable_drawings folder.
The Barkour robots use a custom motor driver that communicates with an on-board computer over EtherCAT:
We have a complete robot assembly guide to build robots from scratch. Take a look at the following assembly guides:
- Barkour robot assembly
- Actuator assembly and setup
- Hip Cover Assembly instruction
- Leg Linkage Assembly Instructions
- Wireless E-stop (STO) assembly
and here for the Bill of Materials (BOM).
If you are curious about the actuated head, you can find the assembly instructions here
The Barkour motor drivers use custom firmware based on the
Pigweed open-source project. You can find the firmware in
the actuator/firmware folder and our firmware guide here.
We also provide a hostside example to control the
actuators via EtherCAT and a guide to
configure EtherCAT.
The gifs below illustrate the hardware capabilities of the Barkour Robot platform:
This release contains the "vB" version of the Barkour Robot. As part of our earlier work, we used an older version of the hardware, called "v0". You can find the assets for the v0 version at the following links:
- CAD,
- simplified CAD for URDF export,
- sim,
- Barkour obstacle course sim,
- Barkour course scoring script.
If you like this work or make use of our designs, please cite:
@article{caluwaerts2023barkour,
title={Barkour: Benchmarking animal-level agility with quadruped robots},
author={Caluwaerts, Ken and Iscen, Atil and Kew, J Chase and Yu, Wenhao and Zhang, Tingnan and Freeman, Daniel and Lee, Kuang-Huei and Lee, Lisa and Saliceti, Stefano and Zhuang, Vincent and others},
journal={arXiv preprint arXiv:2305.14654},
year={2023}
}Copyright 2024 DeepMind Technologies Limited
All software is licensed under the Apache License, Version 2.0 (Apache 2.0); you may not use this file except in compliance with the Apache 2.0 license. You may obtain a copy of the Apache 2.0 license at: https://www.apache.org/licenses/LICENSE-2.0
All other materials are licensed under the Attribution-NonCommercial 4.0 International (CC BY-NC). You may obtain a copy of the CC BY-NC license at: https://creativecommons.org/licenses/by-nc/4.0/legalcode
Unless required by applicable law or agreed to in writing, all software and materials distributed here under the Apache 2.0 or CC BY-NC licenses are distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the licenses for the specific language governing permissions and limitations under those licenses.
This is not an official Google product.
