Accompanying code for the paper Learning Barrier-Certified Polynomial Dynamical Systems for Obstacle Avoidance with Robots by Martin Schonger1*, Hugo T. M. Kussaba1*, Lingyun Chen1, Luis Figueredo2, Abdalla Swikir1, Aude Billard3, and Sami Haddadin1, published in the proceedings of the 2024 International Conference on Robotics and Automation (ICRA 2024). Pre-print available at arXiv:2403.08178.
1Munich Institute of Robotics and Machine Intelligence (MIRMI), Technical University of Munich (TUM), Germany. Abdalla Swikir is also with the Department of Electrical and Electronic Engineering, Omar Al-Mukhtar University (OMU), Albaida, Libya.
2School of Computer Science, University of Nottingham, UK. Luis Figueredo is also an Associated Fellow at the MIRMI, TUM.
3Learning Algorithms and Systems Laboratory, EPFL, Switzerland.
*These authors contributed equally to the paper.
Install MATLAB (tested with R2023a).
Install the required MathWorks toolboxes: Control System Toolbox, Robust Control Toolbox, Optimization Toolbox, Signal Processing Toolbox, Symbolic Math Toolbox, Statistics and Machine Learning Toolbox.
Install the required third party tools: YALMIP, PENLAB+PENBMI, MOSEK, GUROBI, (Optional for plotting: crameri colormaps).
Note Make sure that the non-toolbox paths are before/on top of the toolbox paths.
Run:
git clone https://github.com/martinschonger/abc-ds.git
cd abc-ds
git submodule init
git submodule updateOpen the abc-ds folder in MATLAB.
Configure the desired experiments in main2.m and run this script.
Check the output folder for results and logs.
(Optionally, recreate the plots from the paper with generate_plots.m, and the animations from the video with generate_plots_video.m.)
This software was created as part of Martin Schonger's master's thesis in Computer Science at the Technical University of Munich's (TUM) School of Computation, Information and Technology (CIT).
Copyright © 2023 Martin Schonger
This software is licensed under the GPLv3.