SIPCC_ICRA2021

Mengchao Zhang

03/21/2021

mz17@illinois.edu

This package contains code accompanying the paper "[Semi-Infinite Programming with Complementarity Constraints for Pose Optimization with Pervasive Contact]" by M. Zhang, K. Hauser, in the International Conference on Robotics and Automation (ICRA), 2021.

File structure

├── data                      World, robot, and object files for running the example code
|   └─── ...
├── README.md                 This file
├── graspopt.py               An example program using SIPCC solving grasping planning problem
└── sipcc/                    The core Python module
    ├── geometryopt.py        SIP code for collision-free constraints between geometries, for objects, robot poses, and robot trajectories.
    ├── __init__.py           Tells Python that this is a module
    ├── sip.py                Generic semi-infinite programming code
    ├── sipcc.py              Generic semi-infinite programming with complementarity constraint solving code
    ├── sipcc_problem.py      Generic semi-infinite programming with complementarity constraint problem defination
    ├── graspingopt.py        SIPCC code for grasping planning
    ├── mpcc.py		      MPCC problem solver (currently only has an SNOPT interface, may add our own MPCC solver in the future)
    └── objective.py          Generic objectives for optimization problems

Dependencies

This package requires

Numpy/Scipy
OSQP for quadratic program (QP) solving. OSQP can be installed using

pip install osqp

Other solvers might be supported in the future.

The Klampt 0.8.x Python API (https://klampt.org) to be installed. pip install klampt may work.
cvxopt. Cvxopt can be installed using

pip install cvxopt

sklearn can be installed using

pip install scikit-learn