Implemetation for our paper "SoftMAC: Differentiable Soft Body Simulation with Forecast-based Contact Model and Two-way Coupling with Articulated Rigid Bodies and Clothes" (accepted to IROS 2024).
Soft-rigid and soft-cloth coupling are implemented separately. The underlying cloth simulator DiffClothAI is not open-source yet, but we provide our code for soft-cloth coupling under soft_cloth
. Soft-rigid coupling is already supported in this repository. Please follow the instructions:
conda create -n softmac python=3.8
conda activate softmac
Jade is the underlying articulated rigid body simulator. You can find it here.
pip install -r requirements.txt
pip install -e .
We optimize the action sequences with SoftMAC for each task.
Demo 1: Pour wine
python3 demo_pour.py
Demo 2: Squeeze plasticine
python3 demo_grip.py
Demo 3: Pull door
python3 demo_door.py
Demo 4: Make taco
# run under ./soft_cloth
python3 demo_taco.py
Demo 5: Push towel
# run under ./soft_cloth
python3 demo_hit.py
[2024.1] Add support for velocity control and an example softmac/demo_pour_vel.py
. You can activate this mode by setting rigid_velocity_control=True
in the configuration file.
Feel free to contact me at [email protected] or create a Github issue if you have questions regarding setting up the repository, running examples or adding new examples.
If you find our work useful, please consider citing:
@inproceedings{liu2023softmac,
title={SoftMAC: Differentiable Soft Body Simulation with Forecast-based Contact Model and Two-way Coupling with Articulated Rigid Bodies and Clothes},
author={Liu, Min and Yang, Gang and Luo, Siyuan and Shao, Lin},
booktitle={IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) },
year={2024},
}