tams-group/robio_2021_dexopt

Direct Policy Optimization with Differentiable Physical Consistency for Dexterous Manipulation - ROBIO 2021

Experimental research code used in our ROBIO 2021 paper

Direct Policy Optimization with Differentiable Physical Consistency for Dexterous Manipulation

Dependencies

• TAMSVIZ https://github.com/TAMS-Group/tamsviz
• Shadow and UR robot models

Paper

Direct Policy Optimization with Differentiable Physical Consistency for Dexterous Manipulation

Citation

@INPROCEEDINGS{dexopt,
  author={Ruppel, Philipp and Hendrich, Norman and Zhang, Jianwei},
  booktitle={2021 IEEE International Conference on Robotics and Biomimetics (ROBIO)},
  title={Direct Policy Optimization with Differentiable Physical Consistency for Dexterous Manipulation},
  year={2021},
  volume={},
  number={},
  pages={650-655},
  doi={10.1109/ROBIO54168.2021.9739435}
}