Booster Train

This repository provides a set of reinforcement learning tasks for Booster robots using Isaac Lab.

This repository provides a set of reinforcement learning tasks for Booster robots using Isaac Lab. Currently it includes the fabulous BeyondMimic motion tracking framework adapted to Booster K1 robots. This repository follows the standard Isaac Lab project structure, and is tested with IsaacLab 2.2 and Isaac Sim 5.0.

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

- Clone or copy this project/repository separately from the Isaac Lab installation (i.e. outside the IsaacLab directory): bash git clone https://github.com/BoosterRobotics/boostertrain.git

- Download and install boosterassets: - Clone the boosterassets which contains Booster robot models and motion data. - Install boosterassets python helper following the instructions in the repository.

- Using a python interpreter that has Isaac Lab installed, install the library in editable mode using:

bash # use 'PATHTOisaaclab.sh|bat -p' instead of 'python' if Isaac Lab is not installed in Python venv or conda python -m pip install -e source/boostertrain

- Prepare BeyondMimic motion data: bash # use 'FULLPATHTOisaaclab.sh|bat -p' instead of 'python' if Isaac Lab is not installed in Python venv or conda python scripts/csvtonpz.py --headless --inputfile=<PATHTOBOOSTERASSETS>/motions/K1/<MOTION>.csv --inputfps=<FPS> --outputname=<PATHTOBOOSTERASSETS>/motions/K1/<MOTION>.npz

bash # use 'FULLPATHTOisaaclab.sh|bat -p' instead of 'python' if Isaac Lab is not installed in Python venv or conda python scripts/listenvs.py

bash # use 'FULLPATHTOisaaclab.sh|bat -p' instead of 'python' if Isaac Lab is not installed in Python venv or conda python scripts/rslrl/train.py --task=<TASKNAME> --headless --device cuda:N

- Play a trained policy and export it for deployment: