mano_pybullet's Introduction

mano_pybullet

MANO-based hand models for the PyBullet simulator.

Install

From source code

git clone https://github.com/ikalevatykh/mano_pybullet.git
cd mano_pybullet
pip install -e .

or if you plan to use OpenAI gym environments:

pip install -e .[gym]

Run tests

export MANO_MODELS_DIR='/path/to/mano_v*_*/models/'
pytest

Download MANO models

Register at the MANO website and download the models.
Unzip the file mano_v*_*.zip: unzip mano_v*_*.zip
Set environment variable: export MANO_MODELS_DIR=/path/to/mano_v*_*/models/

Hand models

The package provides MANO-based rigid hand model.

Gym environments

The package also provides several OpenAI gym environments:

HandEnv - base environment with one or two hands on the scene.
HandObjectEnv - base environment with hands and an object on the scene.
HandLiftEnv - environment where the target is to lift an object.

HandPushEnv - environment where the target is to push an object.

Graphical debugging

python -m mano_pybullet.tools.gui_control

usage: GUI debug tool [-h] [--dofs DOFS] [--left-hand] [--right-hand] [--visual-shapes] [--no-visual-shapes] [--self-collisions] [--no-self-collisions]

optional arguments:

-h, --help show help message and exit
--dofs DOFS number of degrees of freedom (20 or 45) [default=20]
--left-hand show left hand
--right-hand show right hand [default]
--visual-shapes show visual shapes [default]
--no-visual-shapes hide visual shapes
--self-collisions enable self collisions [default]
--no-self-collisions disable self collisions

Citation

If you find mano_pybullet useful in your research, please cite the repository using the following BibTeX entry.

@Misc{kalevatykh2020mano_pybullet,
  author =       {Kalevatykh, Igor et al.},
  title =        {mano_pybullet - porting the MANO hand model to the PyBullet simulator},
  howpublished = {Github},
  year =         {2020},
  url =          {https://github.com/ikalevatykh/mano_pybullet}
}