MuJoCo, short for Multi-Joint dynamics with Contact, is a general purpose physics engine that is tailored to scientific use cases such as robotics, biomechanics and machine learning. It was first described in 2012 in a paper by Emanuel Todorov, Tom Erez, and Yuval Tassa, and later commercialized under Roboti LLC.[1] According to a Google Scholar search,[2] as of April 2024 the original publication has been cited 5329 times, and the MuJoCo engine 9250 times.[3] It was described by Zhao and Queralta in their review as one of "the most widely used simulators in the literature".[4]

MuJoCo
Repositoryhttps://github.com/google-deepmind/mujoco
Written inC, C++, Python, C#
LicenseApache-2.0 license

MuJoCo was acquired by Google DeepMind in October 2021 and open-sourced under the Apache 2.0 license in May 2022.[5] Parts of the Deepmind control suite are powered by the MuJoCo engine.[6]

See also

edit

References

edit
  1. ^ Todorov, Emanuel; Erez, Tom; Tassa, Yuval (2012). MuJoCo: A physics engine for model-based control. pp. 5026–5033. doi:10.1109/IROS.2012.6386109. ISBN 978-1-4673-1736-8. Retrieved 2023-12-07.
  2. ^ Mujoco: A physics engine for model-based control search. "Google Scholar". scholar.google.com.au. Retrieved 2024-04-03.
  3. ^ MuJoCo search. "Google Scholar". scholar.google.com.au. Retrieved 2024-04-03.
  4. ^ Zhao, Wenshuai; Queralta, Jorge Pena; Westerlund, Tomi (2020). "Sim-to-Real Transfer in Deep Reinforcement Learning for Robotics: A Survey". 2020 IEEE Symposium Series on Computational Intelligence (SSCI). IEEE. pp. 737–744. arXiv:2009.13303. doi:10.1109/ssci47803.2020.9308468. ISBN 978-1-7281-2547-3.
  5. ^ "Open-sourcing MuJoCo". Google DeepMind. 2022-05-23. Retrieved 2023-12-07.
  6. ^ Tassa, Yuval; Doron, Yotam; Muldal, Alistair; Erez, Tom; Li, Yazhe; Casas, Diego de Las; Budden, David; Abdolmaleki, Abbas; Merel, Josh (2018-01-02), DeepMind Control Suite, arXiv:1801.00690, retrieved 2024-04-03
edit