Skip to content
RobothecaBeta
Log in

unitree_rl_lab

  • Software
  • Humanoids and robot learning
  • Framework

Reinforcement learning environments for training locomotion policies on Unitree Go2, H1 and G1-29dof robots, built on IsaacLab, with a path from MuJoCo simulation to the physical robot.

  • Closed core

Made by Unitree Robotics.

Source on GitHub: github.com/unitreerobotics/unitree_rl_lab (opens in a new tab). Facts read at commit 4960b84 (opens in a new tab).

Licenses

Code
Apache-2.0 (opens in a new tab)

Read by Robotheca from LICENCE (opens in a new tab) (Apache-2.0) at commit 4960b84 (opens in a new tab), on 23 Oct 2026.

Commercial use
Unknown

It needs closed parts under their own terms: read them before any commercial use.

Parts it needs are closed

  • Needs NVIDIA Isaac Sim: Needed for training or simulation; a closed NVIDIA runtime (Omniverse).
  • Needs libunitree_sdk2 (prebuilt): Deploying to a real robot goes through unitree_sdk2, whose core ships as a prebuilt library without source.

Labels

  • Closed coreIt needs closed parts: NVIDIA Isaac Sim and libunitree_sdk2 (prebuilt).

What it does

  • Provides RL training tasks for Unitree Go2, H1 and G1-29dof robots on top of IsaacLab.
  • Training and playback are run through scripts, for example a velocity task for the G1-29dof.
  • Robot model files come from USD assets or URDF files whose directories are set in a config file.
  • Trained policies can first be tested in MuJoCo simulation (sim2sim) before running on hardware.
  • A C++ robot controller is built against unitree_sdk2 to run the policy on the G1-29dof robot.
  • The controller can also drive the physical robot over a named network interface.

Written by Robotheca from the project's README (opens in a new tab) at commit 4960b84.

Getting started

Difficulty
Advanced

Maintenance

Status
Maintained: a commit in the last 12 months
Newest commit
19 Nov 2025 on main
Latest release
None published on GitHub

Checked on GitHub on 23 Oct 2026.

Links

Robots that can run it

No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.

Robotheca lists software as a reference: nothing is hosted or run here, and every fact links to where it was read.