(Geometric) Fabrics

Note on development

This project is still under heavy development and there is a lack of documentation. I, @maxspahn, am committed to improve and maintain that package. However, I rely on people like you to point me to issues and unclear sections of the code. So feel free to leave issues whenever something bugs you.

Fabrics ros-wrapper

The fabrics-ros wrapper will be released very shortly when compatibility is verified.

Geometric Fabrics represent a geometric approach to motion generation for various robot structures. The idea is a next development step after Riemannian Motion Policies and offers increased stability and accessibility.

Holonomic robots	Non-Holonomic robots

Installation

Install the package through pip, using bash pip3 install ".<options>" or from PyPI using bash pip3 install fabrics Options are [agents] and [tutorials]. Those can be installed using pip3 install ".[agents]" pip3 install ".[tutorials]"

Install the package through poetry, using bash poetry install --with <option>

Publications

This repository was used in several publications. The major one being Dynamic Optimization Fabrics for Motion Generation If you are using this software, please cite: bash @misc{https://doi.org/10.48550/arxiv.2205.08454, doi = {10.48550/ARXIV.2205.08454}, url = {https://arxiv.org/abs/2205.08454}, author = {Spahn, Max and Wisse, Martijn and Alonso-Mora, Javier}, keywords = {Robotics (cs.RO), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Dynamic Optimization Fabrics for Motion Generation}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution Share Alike 4.0 International} } Other publications where this repository was used:

https://github.com/maxspahn/optuna_fabrics bash @article{https://doi.org/10.48550/arxiv.2302.06922, doi = {10.48550/ARXIV.2302.06922}, url = {https://arxiv.org/abs/2302.06922}, author = {Spahn, Max and Alonso-Mora, Javier}, keywords = {Robotics (cs.RO), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Autotuning Symbolic Optimization Fabrics for Trajectory Generation}, publisher = {arXiv}, year = {2023}, copyright = {Creative Commons Attribution Share Alike 4.0 International} }

https://github.com/tud-amr/localPlannerBench bash @misc{https://doi.org/10.48550/arxiv.2210.06033, doi = {10.48550/ARXIV.2210.06033}, url = {https://arxiv.org/abs/2210.06033}, author = {Spahn, Max and Salmi, Chadi and Alonso-Mora, Javier}, keywords = {Robotics (cs.RO), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Local Planner Bench: Benchmarking for Local Motion Planning}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution Share Alike 4.0 International} }

Tutorials

This repository contains brief examples corresponding to the theory presented in "Optimization Fabrics" by Ratliff et al. https://arxiv.org/abs/2008.02399. These examples are named according to the naming in that publication. Each script is self-contained and required software is installed using bash pip install ".[tutorials]"

Related works and websites

The work is based on some works by the NVIDIA Research Labs. Below you find a list of all relevant links:

lecture notes

• https://www.nathanratliff.com/pedagogy/mathematics-for-intelligent-systems#lecture6

websites

• https://sites.google.com/nvidia.com/geometric-fabrics

paper

• https://arxiv.org/abs/2010.14750
• https://arxiv.org/abs/2008.02399
• https://arxiv.org/abs/2010.14745
• https://arxiv.org/abs/2010.15676
• https://arxiv.org/abs/1801.02854

videos and talks

• https://www.youtube.com/watch?v=aM9Ha2IawEo
• https://www.youtube.com/watch?v=awiF6JjDEbo
• https://www.youtube.com/watch?v=VsM-kdk74d8