NEW: We now have an analytically differentiable version! See the the notes and the reference MATLAB implementation.
Redmax was also used for differentiable hand simulation, which was presented at RSS 2021. The associated github repository contains the C++ implementation of redmax with Python bindings.
notes.pdf
: An extensive writeup with details on:
matlab-diff
: Object-oriented MATLAB implementation of differentiable redmax
matlab-simple
: Simpler object-oriented MATLAB implementation for getting startedmatlab
: Object-oriented MATLAB implementation with many features, including:
ode45
or euler
c++
: C++ implementation including Projected Block Jacobi PreconditionerACM Transactions on Graphics, 38 (4) 104:1-104:10 (SIGGRAPH), 2019.
Ying Wang, Nicholas J. Weidner, Margaret A. Baxter, Yura Hwang, Danny M. Kaufman, Shinjiro Sueda
@article{Wang2019,
author = {Wang, Ying and Weidner, Nicholas J. and Baxter, Margaret A. and Hwang, Yura and Kaufman, Danny M. and Sueda, Shinjiro},
title = {\textsc{RedMax}: Efficient \& Flexible Approach for Articulated Dynamics},
year = {2019},
issue_date = {July 2019},
publisher = {ACM},
address = {New York, NY, USA},
volume = {38},
number = {4},
issn = {0730-0301},
url = {https://doi.org/10.1145/3306346.3322952},
doi = {10.1145/3306346.3322952},
journal = {{ACM} Trans.\ Graph.},
month = jul,
articleno = {104},
numpages = {10},
keywords = {friction, rigid body dynamics, physical simulation, constraints, contact}
}