This repository is the implentation of the BiConMP solver. The BiConMP is an online whole body trajectory optimizer for legged robots. The trajectory optimizer has generated various dyanmic behaviorus on Quadrupeds (Solo12, AnYmal), Humanoids (Talos)...
The corresponding paper was published in the IEEE Transactions on Robotics and is freely accessible here
- Pinocchio (Version 2.6.4 or greater)
- Crocoddyl (Version 1.8.1 or greater)
- Pybind11
- bullet_utils (needed to run examples and tutorials) : Installation link
- Robot_Properties_Solo (needed to run examples and tutorials) : Installation link
- OsqpEigen (to use OSQP as the solver): Installation Link
- Meshcat (To visualize soltutions) : Installation Link
- Matplotlib (to plot results)
git clone --recurse-submodules git@github.com:machines-in-motion/biconvex_mpc.git
cd biconvex_mpc
mkdir build && cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make install -j16
If you want to use OSQP, the cmake instructions should be modified to the following:
cmake .. -DCMAKE_BUILD_TYPE=Release -DOSQP=TRUE
After building the package add the following lines to your bashrc to ensure that python can find BiconMP.
export PYTHONPATH=$PYTHONPATH:path/to/the/biconvex/build/folder
- Avadesh Meduri
- Paarth Shah
@article{meduri2022biconmp,
title={BiConMP: A Nonlinear Model Predictive Control Framework for Whole Body Motion Planning},
author={Meduri, Avadesh and Shah, Paarth and Viereck, Julian and Khadiv, Majid and Havoutis, Ioannis and Righetti, Ludovic},
journal={IEEE Transactions on Robotics},
pages = {905--922},
volume = {39},
issue = {2},
year={2023}
}
Copyright(c) 2019-2023 New York University
BSD 3-Clause License