Home

Awesome

RF-MPC

Representation-Free Model Predictive Control (RF-MPC) is a MATLAB simulation framework for dynamic legged robots. RF-MPC represents the orientation using the rotation matrix and thus does not have the singularity issue associated with the Euler angles. The linear dynamics on the rotation matrix is derived using variation-based linearization (VBL).

video available at: YouTube Video

Requirement

Basic: MATLAB and MATLAB optimization toolbox

Optional: qpSWIFT (can be obtained from https://github.com/qpSWIFT)

Installation

There is no need to install external packages.

Usage

navigate to the root directory and run the MAIN.m function

MAIN

The Plant

The robot is modeled as a single rigid body (SRB). The SRB dynamics is defined in

...\fcns\dynamics_SRB.m

VBL and vectorization

The code for variation-based linearization and vectorization steps is in

...\fcns_MPC\fcn_get_ABD_eta.m

Quadratic Program (QP)

The code for QP formulation is in

...\fcns_MPC\fcn_get_QP_form_eta.m

The QP could be solved by either the MATLAB QP solver quadprog or a efficient QP solver qpSWIFT (coming soon!)

How to cite

@ARTICLE{9321699,
author={Y. {Ding} and A. {Pandala} and C. {Li} and Y. -H. {Shin} and H. -W. {Park}},
journal={IEEE Transactions on Robotics}, 
title={Representation-Free Model Predictive Control for Dynamic Motions in Quadrupeds}, 
year={2021},
volume={},
number={},
pages={1-18},
doi={10.1109/TRO.2020.3046415}}

References

This code is based on the following publications:

Authors

Yanran Ding - Initial Work/Maintainer

Contributing

For major changes, please open an issue first to discuss what you would like to change.

Acknowledgments