bipedal_locomotion_isaaclab

Overview

This repository is used to train and simulate bipedal robots, such as limxdynamics TRON1. With the help of Isaac Lab, we can train the bipedal robots to walk in different environments, such as flat, rough, and stairs.

Keywords: isaaclab, locomotion, bipedal, pointfoot

Installation

• Install Isaac Lab by following the installation guide. We recommend using the conda installation as it simplifies calling Python scripts from the terminal.

• Clone the repository separately from the Isaac Lab installation (i.e. outside the IsaacLab directory):

# Option 1: HTTPS
git clone https://github.com/Andy-xiong6/bipedal_locomotion_isaaclab.git

# Option 2: SSH
git clone git@github.com:Andy-xiong6/bipedal_locomotion_isaaclab.git

# Enter the repository
conda activate isaaclab
cd bipedal_locomotion_isaaclab

• Using a python interpreter that has Isaac Lab installed, install the library

python -m pip install -e exts/bipedal_locomotion

Set up IDE (Optional)

To setup the IDE, please follow these instructions:

• Run VSCode Tasks, by pressing Ctrl+Shift+P, selecting Tasks: Run Task and running the setup_python_env in the drop down menu. When running this task, you will be prompted to add the absolute path to your Isaac Sim installation.

If everything executes correctly, it should create a file .python.env in the .vscode directory. The file contains the python paths to all the extensions provided by Isaac Sim and Omniverse. This helps in indexing all the python modules for intelligent suggestions while writing code.

Training the bipedal robot agent

• Use the scripts/rsl_rl/train.py script to train the robot directly, specifying the task:

python scripts/rsl_rl/train.py --task=Isaac-PF-Blind-Flat-v0 --headless

• It is recommend to use start.sh script to train the robot, specifying the task in the script:

bash ./start.sh

• The following arguments can be used to customize the training:
  • --headless: Run the simulation in headless mode
  • --num_envs: Number of parallel environments to run
  • --max_iterations: Maximum number of training iterations
  • --save_interval: Interval to save the model
  • --seed: Seed for the random number generator

Playing the trained model

• To play a trained model:

python scripts/rsl_rl/play.py --task=Isaac-PF-Blind-Flat-Play-v0 --checkpoint_path=path/to/checkpoint

• The following arguments can be used to customize the playing:
  • --num_envs: Number of parallel environments to run
  • --headless: Run the simulation in headless mode
  • --checkpoint_path: Path to the checkpoint to load

Running exported model in mujoco (sim2sim)

• After playing the model, the policy has already been saved. You can export the policy to mujoco environment and run it in mujoco @Andy-xiong6/pointfoot-mujoco-sim by using the rl-deploy-with-python repo @Andy-xiong6/rl-deploy-with-python.

• Follwing the instructions to install it properly and replace the model/pointfoot/{Robot Type}/policy/policy.onnx by your trained policy.onnx.

Running exported model in real robot (sim2real)

• TODO

Video Demonstration

Simulation in Isaac Lab

• Pointfoot Blind Flat:

flat-play.mov

• Pointfoot Blind Rough:

rough-play.mov

• Pointfoot Blind Stairs:

stairs-play.mov

Simulation in Mujoco

• Pointfoot Blind Flat:

mujoco.mov

Troubleshooting

Pylance Missing Indexing of Extensions

In some VsCode versions, the indexing of part of the extensions is missing. In this case, add the path to your extension in .vscode/settings.json under the key "python.analysis.extraPaths".

{
    "python.analysis.extraPaths": [
        "<path-to-ext-repo>/exts/bipedal_locomotion"
    ]
}

Pylance Crash

If you encounter a crash in pylance, it is probable that too many files are indexed and you run out of memory. A possible solution is to exclude some of omniverse packages that are not used in your project. To do so, modify .vscode/settings.json and comment out packages under the key "python.analysis.extraPaths" Some examples of packages that can likely be excluded are:

"<path-to-isaac-sim>/extscache/omni.anim.*"         // Animation packages
"<path-to-isaac-sim>/extscache/omni.kit.*"          // Kit UI tools
"<path-to-isaac-sim>/extscache/omni.graph.*"        // Graph UI tools
"<path-to-isaac-sim>/extscache/omni.services.*"     // Services tools
...

Acknowledgements

This project uses the following open-source libraries:

• IsaacLabExtensionTemplate

This project is inspired by the work of H. Wang, H. Luo, W. Zhang, and H. Chen (2024), which laid the groundwork for the idea behind this repository.

This project is part of the Bipedal Robot Control project led by Prof. Wei Zhang at SUSTech (Southern University of Science and Technology).

Special thanks to all the contributors for their hard work and dedication:

• Prof. Wei Zhang
• Hongwei Xiong
• Haoxiang Luo
• Junde Guo
• Bobin Wang
• Haojia Zhang

Special thanks to CLEAR Lab for their technical and financial support during the development of this project.