https://github.com/trajectory-invariants/invariants_py
Calculate invariant trajectory representations from trajectory data and generate new trajectories from the invariants.
https://github.com/trajectory-invariants/invariants_py
casadi coordinate-invariant differential-geometry frenet-serret geometric-optimal-control invariance invariant invariants kinematics optimal-control optimization python robotics screw-theory trajectory-analysis trajectory-generation trajectory-optimization trajectory-planning trajectory-representation
Last synced: about 1 month ago
JSON representation
Calculate invariant trajectory representations from trajectory data and generate new trajectories from the invariants.
- Host: GitHub
- URL: https://github.com/trajectory-invariants/invariants_py
- Owner: trajectory-invariants
- License: mit
- Created: 2023-09-20T14:20:01.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-10-29T14:42:24.000Z (7 months ago)
- Last Synced: 2024-10-29T15:00:27.802Z (7 months ago)
- Topics: casadi, coordinate-invariant, differential-geometry, frenet-serret, geometric-optimal-control, invariance, invariant, invariants, kinematics, optimal-control, optimization, python, robotics, screw-theory, trajectory-analysis, trajectory-generation, trajectory-optimization, trajectory-planning, trajectory-representation
- Language: Jupyter Notebook
- Homepage: https://trajectory-invariants.github.io/
- Size: 2.24 MB
- Stars: 4
- Watchers: 1
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
- Citation: CITATION.bib
Awesome Lists containing this project
README
[](https://opensource.org/licenses/MIT)
`invariants-py` is a Python library to robustly calculate coordinate-invariant trajectory representations using geometric optimal control.
It also supports trajectory generation under user-specified trajectory constraints starting from invariant representations.
More information can be found on the documentation website: https://trajectory-invariants.github.io
## Features
The main features are:
- Calculation of invariant descriptors for trajectories.
- Fast trajectory adaptation starting from the invariant descriptors.
Invariant trajectory representations find their application in trajectory analysis, trajectory segmentation, recognition and generalization.
## Installation
The package can be installed from PyPI or from source.
### Prerequisites
The package requires Python 3.8 or higher.
### 1. Installation from PyPI
This installation option is recommended if you only want to use the package and do not plan to modify the source code.
Upgrade your version of `pip`:
```shell
pip install --upgrade pip
```
Install the package:
```shell
pip install invariants-py
```
### 2. Installation from source
This installation option is recommended if you plan to modify or experiment with the source code.
Clone (or download) the `invariants-py` repository:
```shell
git clone https://github.com/trajectory-invariants/invariants_py.git
```
Navigate to the cloned repository:
```shell
cd invariants_py
```
Upgrade your version of `pip`:
```shell
pip install --upgrade pip
```
Install the package:
```shell
pip install -e .
```
## Getting started
Basic examples are provided to get you started in the [examples folder](./examples/).
More detailed examples and tutorials can be found on the [documentation website](https://trajectory-invariants.github.io/docs/python/).
## Speed up using Fatrop
To speed up the solution of the optimal control problems, you can optionally install the [fatrop solver](https://gitlab.kuleuven.be/robotgenskill/fatrop/fatrop). The instructions are available on [this page](https://trajectory-invariants.github.io/docs/python/installation/installation-fatrop/) of the documentation website.
## Roadmap
The following features are planned for future releases:
- Support for more types of invariant representations (e.g. screw invariants, global invariants, ...).
- Support for more types of constraints in the trajectory generation.
- Benchmarking between different invariant representations in terms of robustness, computational efficiency, and generalizability.
## Contributing
We welcome contributions to this repository, for example in the form of pull requests.
## Contributors
We wish to thank the following people for their contributions to an early version of the software: Victor van Wymeersch, Zeno Gillis, Toon Daemen, Glenn Maes, Ali Mousavi, Lander Vanroye
## Support
For questions, bugs, feature requests, etc., please open [an issue on this repository](https://github.com/trajectory-invariants/invariants_py/issues).
## License
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
## Citing
If you use this package in your research, please cite the following paper:
```
@article{vochten2023invariant,
title={Invariant Descriptors of Motion and Force Trajectories for Interpreting Object Manipulation Tasks in Contact},
author={Vochten, Maxim and Mohammadi, Ali Mousavi and Verduyn, Arno and De Laet, Tinne and Aertbeli{\"e}n, Erwin and De Schutter, Joris},
journal={IEEE Transactions on Robotics},
year={2023},
volume={39},
number={6},
pages={4892-4912},
doi={10.1109/TRO.2023.3309230}}
```