Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Mk1 Humanoid Robot - Isaac Lab Environment

This repository contains all the necessary files to simulate and train the Mk1 humanoid robot, as featured in my YouTube series. This project provides the simulation assets (URDF, USD, meshes), the complete Isaac Lab environment code, and a pre-trained policy for active balancing.

Watch the full story of this robot's development on YouTube


Prerequisites

Before you begin, ensure you have NVIDIA Isaac Lab installed and working correctly. This project is a custom environment and relies on that framework.


Installation

Follow these steps carefully to set up the environment.

1. Place the Project Folder

Copy the entire robotmk1 project folder into your Isaac Lab tasks directory. The recommended location is:

<path_to_isaaclab>/source/isaaclab_tasks/isaaclab_tasks/manager_based/ .../manager_based/ │ └── robotmk1/ <-- YOUR PROJECT FOLDER GOES HERE ├── init.py ├── robotmk1_env_active_balance_cfg.py └── ... (other files) code Code

2. Configure the Environment Path

Important

This is the most critical step. If this path is wrong, the environment will not be found.

Open the __init__.py file inside your new robotmk1 folder. You must edit the env_cfg_entry_point variable to match the location from Step 1.

  • Find this line:
    "env_cfg_entry_point": "CHANGE_THIS_TO_YOUR_ENV_CONFIG_PATH.robotmk1_env_active_balance_cfg:RobotMk1ActiveBalanceEnvCfg",
  • Change it to the correct Python import path:
    "env_cfg_entry_point": "isaaclab_tasks.manager_based.robotmk1.robotmk1_env_active_balance_cfg:RobotMk1ActiveBalanceEnvCfg",

3. Configure the Robot Asset Path

Open the robotmk1_env_active_balance_cfg.py file. You must edit the robot_usd_path variable to point to the absolute path of the mk1.usd file in this repository.

  • Find this line:
    robot_usd_path = r"PATH TO THE ROBOT USD"
  • Change it to the full path on your system, for example:
    robot_usd_path = r"C:\Users\YourName\Documents\GitHub\robotmk1\mk1.usd"

Usage

All commands should be run from the root of your Isaac Lab repository.

Running the Pre-trained Policy

This repository includes a working policy for active balancing. To see it in action, run the following command. Make sure to replace the placeholder with the correct path to the included .pth file.

Project Structure

  • /Autocad Inventor: Contains all files related to the Autocad Inventor CAD.
  • /Solidworks: Contains all files related to the Solidworks CAD.
  • /isaacsim: Contains all files related to the Isaac Lab simulation.
    • /isaacsim/mk1: This is the core environment folder. It contains all the Python code that needs to be copied into the Isaac Lab tasks directory.
      • /agents: Holds the RL Games configuration (.yaml) for the training agent.
      • /mdp: Holds the custom Python files (rewards.py, observations.py, etc.) that define the environment's logic.
    • /isaacsim/trained weight: Contains the pre-trained balancing policy (.pth file).
  • /meshes: Contains all the individual .stl part files for the robot's geometry.
  • /urdf: Contains the mk1.urdf file defining the robot's joints and links.
  • /usd: Contains the mk1.usd file used for rendering in Isaac Sim.

License

This project is released into the public domain under The Unlicense.

This means you are free to copy, modify, distribute, and use these files for any purpose, without any conditions.

While not required, a link back to my YouTube channel is always appreciated if you find this project helpful!


Created by Wirdegeist

About

This repository contains the simulation assets for the Mk1 humanoid robot featured in my YouTube video.

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages