Berkeley Humanoid

Low-cost mid-scale research biped for learning-based control, 0.85 m and 16 kg with 12 leg DOF driven by in-house actuators.

Frames · Berkeley Humanoid

4 items · credit and license on each

Studio render of the published MJCF, keyframe pose from the model. Geometry only; colours are the model's own materials.Rendered by Hyperspawn from mujoco_menagerie/berkeley_humanoid (BSD-3-Clause) · Hyperspawn render · source
Side view of the same MJCF render, keyframe pose from the model.Rendered by Hyperspawn from mujoco_menagerie/berkeley_humanoid (BSD-3-Clause) · Hyperspawn render · source
4 s turntable of the MJCF studio render.Rendered by Hyperspawn from mujoco_menagerie/berkeley_humanoid (BSD-3-Clause) · Hyperspawn render · source
Berkeley Humanoid: A Research Platform for Learning-based ControlHybrid Robotics · official embed · source

Provenance

Model: Berkeley Humanoid
Repo: google-deepmind/mujoco_menagerie
Commit: c96a32d28fb5
Entry: berkeley_humanoid/scene.xml
Files: 15 · 26 MB, fetched from raw.githubusercontent.com, cached in your browser
License: BSD-3-Clause
Engine: MuJoCo 3.14.0, official WebAssembly build · three.js renderer
Controller: PD hold on the stand keyframe with an ankle-strategy tilt term; not a learned policy
Note: The original Berkeley Humanoid, not Humanoid Lite.

Policies listed for Berkeley Humanoid

No published policy we can replay yet.

Nothing is playing a trained controller in the viewport above. What you see is a PD hold on the stand pose, computed in your tab. A policy plays here when a checkpoint is published with an obs/action contract we can reproduce exactly in MuJoCo, and it passes a 30 s walk test in the tab.

Set up a training run

Train on GPUs

mjlab ships no stock task for Berkeley Humanoid. The training page gives you the pinned model and the steps to write a robot config.

Build the job →