fan-ziqi/robot_lab

Health v2 · maintenance only

Adoption not part of the score

1937 stars · 240 forks observed · 2026-10-14

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

robot_lab is a reinforcement learning extension library for robots built on top of NVIDIA IsaacLab, providing preconfigured locomotion environments for quadrupeds, wheeled-legged, and humanoid robots. It lets researchers develop and train RL policies in an isolated environment outside the core Isaac Lab repository.

Use cases

• train RL locomotion policies for quadruped robots like Unitree Go2
• simulate legged robots in Isaac Sim with IsaacLab
• develop custom robot RL environments outside the IsaacLab core repo
• train velocity-tracking rough terrain policies for ANYmal or Unitree robots
• learn wheeled-legged robot control with reinforcement learning
• experiment with humanoid robot RL training

When to choose

• you already use IsaacLab/Isaac Sim and want ready-made robot RL environments
• you need to train locomotion policies for popular quadruped models
• you want an isolated extension environment for custom IsaacLab tasks

When to avoid

• you need real-robot deployment code without simulation
• you don't have access to NVIDIA Isaac Sim / a suitable GPU
• you want a general-purpose RL library unrelated to robotics

Facets

library · maturity active

reinforcement-learning simulation machine-learning robotics robotics reinforcement-learning simulation machine-learning windows python isaac-lab isaac-sim quadruped-robots legged-locomotion sim-to-real robot-learning linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="fan-ziqi/robot_lab")