| Name | jaxsim JSON |
| Version |
0.5.0
JSON |
| download |
| home_page | None |
| Summary | A differentiable physics engine and multibody dynamics library for control and robot learning. |
| upload_time | 2024-11-15 17:47:57 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.10 |
| license | BSD 3-Clause License Copyright (c) 2022, Artificial and Mechanical Intelligence All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. 3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| keywords |
physics
physics engine
jax
rigid body dynamics
featherstone
reinforcement learning
robot
robotics
sdf
urdf
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# JaxSim
JaxSim is a **differentiable physics engine** and **multibody dynamics library** designed for applications in control and robot learning, implemented with JAX.
Its design facilitates research and accelerates prototyping in the intersection of robotics and artificial intelligence.
<div align="center">
<br/>
<table>
<tr>
<th><img src="https://github.com/user-attachments/assets/115b1c1c-6ae5-4c59-92e0-1be13ba954db" width="250"></th>
<th><img src="https://github.com/user-attachments/assets/f9661fae-9a85-41dd-9a58-218758ec8c9c" width="250"></th>
<th><img src="https://github.com/user-attachments/assets/ae8adadf-3bca-47b8-97ca-3a9273633d60" width="250"></th>
</tr>
</table>
<br/>
</div>
## Features
- Physics engine in reduced coordinates supporting fixed-base and floating-base robots.
- Multibody dynamics library providing all the necessary components for developing model-based control algorithms.
- Completely developed in Python with [`google/jax`][jax] following a functional programming paradigm.
- Transparent support for running on CPUs, GPUs, and TPUs.
- Full support for JIT compilation for increased performance.
- Full support for automatic vectorization for massive parallelization of open-loop and closed-loop architectures.
- Support for SDF models and, upon conversion with [sdformat][sdformat], URDF models.
- Visualization based on the [passive viewer][passive_viewer_mujoco] of Mujoco.
### JaxSim as a simulator
- Wide range of fixed-step explicit Runge-Kutta integrators.
- Support for variable-step integrators implemented as embedded Runge-Kutta schemes.
- Improved stability by optionally integrating the base orientation on the $\text{SO}(3)$ manifold.
- Soft contacts model supporting full friction cone and sticking-slipping transition.
- Collision detection between points rigidly attached to bodies and uneven ground surfaces.
### JaxSim as a multibody dynamics library
- Provides rigid body dynamics algorithms (RBDAs) like RNEA, ABA, CRBA, and Jacobians.
- Provides all the quantities included in the Euler-Poincarè formulation of the equations of motion.
- Supports body-fixed, inertial-fixed, and mixed [velocity representations][notation].
- Exposes all the necessary quantities to develop controllers in centroidal coordinates.
- Supports running open-loop and full closed-loop control architectures on hardware accelerators.
### JaxSim for robot learning
- Being developed with JAX, all the RBDAs support automatic differentiation both in forward and reverse modes.
- Support for automatically differentiating against kinematics and dynamics parameters.
- All fixed-step integrators are forward and reverse differentiable.
- All variable-step integrators are forward differentiable.
- Ideal for sampling synthetic data for reinforcement learning (RL).
- Ideal for designing physics-informed neural networks (PINNs) with loss functions requiring model-based quantities.
- Ideal for combining model-based control with learning-based components.
[jax]: https://github.com/google/jax/
[sdformat]: https://github.com/gazebosim/sdformat
[notation]: https://research.tue.nl/en/publications/multibody-dynamics-notation-version-2
[passive_viewer_mujoco]: https://mujoco.readthedocs.io/en/stable/python.html#passive-viewer
> [!WARNING]
> This project is still experimental, APIs could change between releases without notice.
> [!NOTE]
> JaxSim currently focuses on locomotion applications.
> Only contacts between bodies and smooth ground surfaces are supported.
## Documentation
The JaxSim API documentation is available at [jaxsim.readthedocs.io][readthedocs].
[readthedocs]: https://jaxsim.readthedocs.io/
## Installation
<details>
<summary>With conda</summary>
You can install the project using [`conda`][conda] as follows:
```bash
conda install jaxsim -c conda-forge
```
You can enforce GPU support, if needed, by also specifying `"jaxlib = * = *cuda*"`.
</details>
<details>
<summary>With pip</summary>
You can install the project using [`pypa/pip`][pip], preferably in a [virtual environment][venv], as follows:
```bash
pip install jaxsim
```
Check [`pyproject.toml`](pyproject.toml) for the complete list of optional dependencies.
You can obtain a full installation using `jaxsim[all]`.
If you need GPU support, follow the official [installation instructions][jax_gpu] of JAX.
</details>
<details>
<summary>Contributors installation</summary>
If you want to contribute to the project, we recommend creating the following `jaxsim` conda environment first:
```bash
conda env create -f environment.yml
```
Then, activate the environment and install the project in editable mode:
```bash
conda activate jaxsim
pip install --no-deps -e .
```
</details>
[conda]: https://anaconda.org/
[pip]: https://github.com/pypa/pip/
[venv]: https://docs.python.org/3/tutorial/venv.html
[jax_gpu]: https://github.com/google/jax/#installation
## Overview
<details>
<summary>Structure of the Python package</summary>
```
# tree -L 2 -I "__pycache__" -I "__init__*" -I "__main__*" src/jaxsim
src/jaxsim
|-- api..........................# Package containing the main functional APIs.
| |-- com.py...................# |-- APIs for computing quantities related to the center of mass.
| |-- common.py................# |-- Common utilities used in the current package.
| |-- contact.py...............# |-- APIs for computing quantities related to the collidable points.
| |-- data.py..................# |-- Class storing the data of a simulated model.
| |-- frame.py.................# |-- APIs for computing quantities related to additional frames.
| |-- joint.py.................# |-- APIs for computing quantities related to the joints.
| |-- kin_dyn_parameters.py....# |-- Class storing kinematic and dynamic parameters of a model.
| |-- link.py..................# |-- APIs for computing quantities related to the links.
| |-- model.py.................# |-- Class defining a simulated model and APIs for computing related quantities.
| |-- ode.py...................# |-- APIs for computing quantities related to the system dynamics.
| |-- ode_data.py..............# |-- Set of classes to store the data of the system dynamics.
| `-- references.py............# `-- Helper class to create references (link forces and joint torques).
|-- exceptions.py................# Module containing functions to raise exceptions from JIT-compiled functions.
|-- integrators..................# Package containing the integrators used to simulate the system dynamics.
| |-- common.py................# |-- Common utilities used in the current package.
| |-- fixed_step.py............# |-- Fixed-step integrators (explicit Runge-Kutta schemes).
| `-- variable_step.py.........# `-- Variable-step integrators (embedded Runge-Kutta schemes).
|-- logging.py...................# Module containing logging utilities.
|-- math.........................# Package containing mathematical utilities.
| |-- adjoint.py...............# |-- APIs for creating and manipulating 6D transformations.
| |-- cross.py.................# |-- APIs for computing cross products of 6D quantities.
| |-- inertia.py...............# |-- APIs for creating and manipulating 6D inertia matrices.
| |-- joint_model.py...........# |-- APIs defining the supported joint model and the corresponding transformations.
| |-- quaternion.py............# |-- APIs for creating and manipulating quaternions.
| |-- rotation.py..............# |-- APIs for creating and manipulating rotation matrices.
| |-- skew.py..................# |-- APIs for creating and manipulating skew-symmetric matrices.
| `-- transform.py.............# `-- APIs for creating and manipulating homogeneous transformations.
|-- mujoco.......................# Package containing utilities to interact with the Mujoco passive viewer.
| |-- loaders.py...............# |-- Utilities for converting JaxSim models to Mujoco models.
| |-- model.py.................# |-- Class providing high-level methods to compute quantities using Mujoco.
| `-- visualizer.py............# `-- Class that simplifies opening the passive viewer and recording videos.
|-- parsers......................# Package containing utilities to parse model descriptions (SDF and URDF models).
| |-- descriptions/............# |-- Package containing the intermediate representation of a model description.
| |-- kinematic_graph.py.......# |-- Definition of the kinematic graph associated with a parsed model description.
| `-- rod/.....................# `-- Package to create the intermediate representation from model descriptions using ROD.
|-- rbda.........................# Package containing the low-level rigid body dynamics algorithms.
| |-- aba.py...................# |-- The Articulated Body Algorithm.
| |-- collidable_points.py.....# |-- Kinematics of collidable points.
| |-- contacts/................# |-- Package containing the supported contact models.
| |-- crba.py..................# |-- The Composite Rigid Body Algorithm.
| |-- forward_kinematics.py....# |-- Forward kinematics of the model.
| |-- jacobian.py..............# |-- Full Jacobian and full Jacobian derivative.
| |-- rnea.py..................# |-- The Recursive Newton-Euler Algorithm.
| `-- utils.py.................# `-- Common utilities used in the current package.
|-- terrain......................# Package containing resources to specify the terrain.
| `-- terrain.py...............# `-- Classes defining the supported terrains.
|-- typing.py....................# Module containing type hints.
`-- utils........................# Package of common utilities.
|-- jaxsim_dataclass.py......# |-- Utilities to operate on pytree dataclasses.
|-- tracing.py...............# |-- Utilities to use when JAX is tracing functions.
`-- wrappers.py..............# `-- Utilities to wrap objects for specific use cases on pytree dataclass attributes.
```
</details>
## Credits
The RBDAs are based on the theory of the [Rigid Body Dynamics Algorithms][RBDA]
book by Roy Featherstone.
The algorithms and some simulation features were inspired by its accompanying [code][spatial_v2].
[RBDA]: https://link.springer.com/book/10.1007/978-1-4899-7560-7
[spatial_v2]: http://royfeatherstone.org/spatial/index.html#spatial-software
The development of JaxSim started in late 2021, inspired by early versions of [`google/brax`][brax].
At that time, Brax was implemented in maximal coordinates, and we wanted a physics engine in reduced coordinates.
We are grateful to the Brax team for their work and showing the potential of [JAX][jax] in this field.
Brax v2 was later implemented reduced coordinates, following an approach comparable to JaxSim.
The development then shifted to [MJX][mjx], which today provides a JAX-based implementation of the Mujoco APIs.
The main differences between MJX/Brax and JaxSim are as follows:
- JaxSim supports out-of-the-box all SDF models with [Pose Frame Semantics][PFS].
- JaxSim only supports collisions between points rigidly attached to bodies and a compliant ground surface.
Our contact model requires careful tuning of its spring-damper parameters, but being an instantaneous
function of the state $(\mathbf{q}, \boldsymbol{\nu})$, it doesn't require running any optimization algorithm
when stepping the simulation forward.
- JaxSim mitigates the stiffness of the contact-aware system dynamics by providing variable-step integrators.
[brax]: https://github.com/google/brax
[mjx]: https://mujoco.readthedocs.io/en/3.0.0/mjx.html
[PFS]: http://sdformat.org/tutorials?tut=pose_frame_semantics
## Contributing
We welcome contributions from the community.
Please read the [contributing guide](./CONTRIBUTING.md) to get started.
## Citing
```bibtex
@software{ferigo_jaxsim_2022,
author = {Diego Ferigo and Filippo Luca Ferretti and Silvio Traversaro and Daniele Pucci},
title = {{JaxSim}: A Differentiable Physics Engine and Multibody Dynamics Library for Control and Robot Learning},
url = {http://github.com/ami-iit/jaxsim},
year = {2022},
}
```
Theoretical aspects of JaxSim are based on Chapters 7 and 8 of the following Ph.D. thesis:
```bibtex
@phdthesis{ferigo_phd_thesis_2022,
title = {Simulation Architectures for Reinforcement Learning applied to Robotics},
author = {Diego Ferigo},
school = {University of Manchester},
type = {PhD Thesis},
month = {July},
year = {2022},
}
```
## People
| Authors | Maintainers |
|:------:|:-----------:|
| [<img src="https://avatars.githubusercontent.com/u/469199?v=4" width="40">][df] [<img src="https://avatars.githubusercontent.com/u/102977828?v=4" width="40">][ff] | [<img src="https://avatars.githubusercontent.com/u/102977828?v=4" width="40">][ff] [<img src="https://avatars.githubusercontent.com/u/57228872?v=4" width="40">][ac] |
[df]: https://github.com/diegoferigo
[ff]: https://github.com/flferretti
[ac]: https://github.com/xela-95
## License
[BSD3](https://choosealicense.com/licenses/bsd-3-clause/)
Raw data
{
"_id": null,
"home_page": null,
"name": "jaxsim",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.10",
"maintainer_email": "Filippo Luca Ferretti <filippo.ferretti@iit.it>, Alessandro Croci <alessandro.croci@iit.it>",
"keywords": "physics, physics engine, jax, rigid body dynamics, featherstone, reinforcement learning, robot, robotics, sdf, urdf",
"author": null,
"author_email": "Diego Ferigo <dgferigo@gmail.com>, Filippo Luca Ferretti <filippoluca.ferretti@outlook.com>",
"download_url": "https://files.pythonhosted.org/packages/e0/01/d88d7844ac5f5fb93510ce4de285b18fbe567c32a058ee32d3ebc8cc06e6/jaxsim-0.5.0.tar.gz",
"platform": null,
"description": "# JaxSim\n\nJaxSim is a **differentiable physics engine** and **multibody dynamics library** designed for applications in control and robot learning, implemented with JAX.\n\nIts design facilitates research and accelerates prototyping in the intersection of robotics and artificial intelligence.\n\n<div align=\"center\">\n<br/>\n<table>\n <tr>\n <th><img src=\"https://github.com/user-attachments/assets/115b1c1c-6ae5-4c59-92e0-1be13ba954db\" width=\"250\"></th>\n <th><img src=\"https://github.com/user-attachments/assets/f9661fae-9a85-41dd-9a58-218758ec8c9c\" width=\"250\"></th>\n <th><img src=\"https://github.com/user-attachments/assets/ae8adadf-3bca-47b8-97ca-3a9273633d60\" width=\"250\"></th>\n </tr>\n</table>\n<br/>\n</div>\n\n## Features\n\n- Physics engine in reduced coordinates supporting fixed-base and floating-base robots.\n- Multibody dynamics library providing all the necessary components for developing model-based control algorithms.\n- Completely developed in Python with [`google/jax`][jax] following a functional programming paradigm.\n- Transparent support for running on CPUs, GPUs, and TPUs.\n- Full support for JIT compilation for increased performance.\n- Full support for automatic vectorization for massive parallelization of open-loop and closed-loop architectures.\n- Support for SDF models and, upon conversion with [sdformat][sdformat], URDF models.\n- Visualization based on the [passive viewer][passive_viewer_mujoco] of Mujoco.\n\n### JaxSim as a simulator\n\n- Wide range of fixed-step explicit Runge-Kutta integrators.\n- Support for variable-step integrators implemented as embedded Runge-Kutta schemes.\n- Improved stability by optionally integrating the base orientation on the $\\text{SO}(3)$ manifold.\n- Soft contacts model supporting full friction cone and sticking-slipping transition.\n- Collision detection between points rigidly attached to bodies and uneven ground surfaces.\n\n### JaxSim as a multibody dynamics library\n\n- Provides rigid body dynamics algorithms (RBDAs) like RNEA, ABA, CRBA, and Jacobians.\n- Provides all the quantities included in the Euler-Poincar\u00e8 formulation of the equations of motion.\n- Supports body-fixed, inertial-fixed, and mixed [velocity representations][notation].\n- Exposes all the necessary quantities to develop controllers in centroidal coordinates.\n- Supports running open-loop and full closed-loop control architectures on hardware accelerators.\n\n### JaxSim for robot learning\n\n- Being developed with JAX, all the RBDAs support automatic differentiation both in forward and reverse modes.\n- Support for automatically differentiating against kinematics and dynamics parameters.\n- All fixed-step integrators are forward and reverse differentiable.\n- All variable-step integrators are forward differentiable.\n- Ideal for sampling synthetic data for reinforcement learning (RL).\n- Ideal for designing physics-informed neural networks (PINNs) with loss functions requiring model-based quantities.\n- Ideal for combining model-based control with learning-based components.\n\n[jax]: https://github.com/google/jax/\n[sdformat]: https://github.com/gazebosim/sdformat\n[notation]: https://research.tue.nl/en/publications/multibody-dynamics-notation-version-2\n[passive_viewer_mujoco]: https://mujoco.readthedocs.io/en/stable/python.html#passive-viewer\n\n> [!WARNING]\n> This project is still experimental, APIs could change between releases without notice.\n\n> [!NOTE]\n> JaxSim currently focuses on locomotion applications.\n> Only contacts between bodies and smooth ground surfaces are supported.\n\n## Documentation\n\nThe JaxSim API documentation is available at [jaxsim.readthedocs.io][readthedocs].\n\n[readthedocs]: https://jaxsim.readthedocs.io/\n\n## Installation\n\n<details>\n<summary>With conda</summary>\n\nYou can install the project using [`conda`][conda] as follows:\n\n```bash\nconda install jaxsim -c conda-forge\n```\n\nYou can enforce GPU support, if needed, by also specifying `\"jaxlib = * = *cuda*\"`.\n\n</details>\n\n<details>\n<summary>With pip</summary>\n\nYou can install the project using [`pypa/pip`][pip], preferably in a [virtual environment][venv], as follows:\n\n```bash\npip install jaxsim\n```\n\nCheck [`pyproject.toml`](pyproject.toml) for the complete list of optional dependencies.\nYou can obtain a full installation using `jaxsim[all]`.\n\nIf you need GPU support, follow the official [installation instructions][jax_gpu] of JAX.\n\n</details>\n\n<details>\n<summary>Contributors installation</summary>\n\nIf you want to contribute to the project, we recommend creating the following `jaxsim` conda environment first:\n\n```bash\nconda env create -f environment.yml\n```\n\nThen, activate the environment and install the project in editable mode:\n\n```bash\nconda activate jaxsim\npip install --no-deps -e .\n```\n\n</details>\n\n[conda]: https://anaconda.org/\n[pip]: https://github.com/pypa/pip/\n[venv]: https://docs.python.org/3/tutorial/venv.html\n[jax_gpu]: https://github.com/google/jax/#installation\n\n## Overview\n\n<details>\n<summary>Structure of the Python package</summary>\n\n```\n# tree -L 2 -I \"__pycache__\" -I \"__init__*\" -I \"__main__*\" src/jaxsim\n\nsrc/jaxsim\n|-- api..........................# Package containing the main functional APIs.\n| |-- com.py...................# |-- APIs for computing quantities related to the center of mass.\n| |-- common.py................# |-- Common utilities used in the current package.\n| |-- contact.py...............# |-- APIs for computing quantities related to the collidable points.\n| |-- data.py..................# |-- Class storing the data of a simulated model.\n| |-- frame.py.................# |-- APIs for computing quantities related to additional frames.\n| |-- joint.py.................# |-- APIs for computing quantities related to the joints.\n| |-- kin_dyn_parameters.py....# |-- Class storing kinematic and dynamic parameters of a model.\n| |-- link.py..................# |-- APIs for computing quantities related to the links.\n| |-- model.py.................# |-- Class defining a simulated model and APIs for computing related quantities.\n| |-- ode.py...................# |-- APIs for computing quantities related to the system dynamics.\n| |-- ode_data.py..............# |-- Set of classes to store the data of the system dynamics.\n| `-- references.py............# `-- Helper class to create references (link forces and joint torques).\n|-- exceptions.py................# Module containing functions to raise exceptions from JIT-compiled functions.\n|-- integrators..................# Package containing the integrators used to simulate the system dynamics.\n| |-- common.py................# |-- Common utilities used in the current package.\n| |-- fixed_step.py............# |-- Fixed-step integrators (explicit Runge-Kutta schemes).\n| `-- variable_step.py.........# `-- Variable-step integrators (embedded Runge-Kutta schemes).\n|-- logging.py...................# Module containing logging utilities.\n|-- math.........................# Package containing mathematical utilities.\n| |-- adjoint.py...............# |-- APIs for creating and manipulating 6D transformations.\n| |-- cross.py.................# |-- APIs for computing cross products of 6D quantities.\n| |-- inertia.py...............# |-- APIs for creating and manipulating 6D inertia matrices.\n| |-- joint_model.py...........# |-- APIs defining the supported joint model and the corresponding transformations.\n| |-- quaternion.py............# |-- APIs for creating and manipulating quaternions.\n| |-- rotation.py..............# |-- APIs for creating and manipulating rotation matrices.\n| |-- skew.py..................# |-- APIs for creating and manipulating skew-symmetric matrices.\n| `-- transform.py.............# `-- APIs for creating and manipulating homogeneous transformations.\n|-- mujoco.......................# Package containing utilities to interact with the Mujoco passive viewer.\n| |-- loaders.py...............# |-- Utilities for converting JaxSim models to Mujoco models.\n| |-- model.py.................# |-- Class providing high-level methods to compute quantities using Mujoco.\n| `-- visualizer.py............# `-- Class that simplifies opening the passive viewer and recording videos.\n|-- parsers......................# Package containing utilities to parse model descriptions (SDF and URDF models).\n| |-- descriptions/............# |-- Package containing the intermediate representation of a model description.\n| |-- kinematic_graph.py.......# |-- Definition of the kinematic graph associated with a parsed model description.\n| `-- rod/.....................# `-- Package to create the intermediate representation from model descriptions using ROD.\n|-- rbda.........................# Package containing the low-level rigid body dynamics algorithms.\n| |-- aba.py...................# |-- The Articulated Body Algorithm.\n| |-- collidable_points.py.....# |-- Kinematics of collidable points.\n| |-- contacts/................# |-- Package containing the supported contact models.\n| |-- crba.py..................# |-- The Composite Rigid Body Algorithm.\n| |-- forward_kinematics.py....# |-- Forward kinematics of the model.\n| |-- jacobian.py..............# |-- Full Jacobian and full Jacobian derivative.\n| |-- rnea.py..................# |-- The Recursive Newton-Euler Algorithm.\n| `-- utils.py.................# `-- Common utilities used in the current package.\n|-- terrain......................# Package containing resources to specify the terrain.\n| `-- terrain.py...............# `-- Classes defining the supported terrains.\n|-- typing.py....................# Module containing type hints.\n`-- utils........................# Package of common utilities.\n |-- jaxsim_dataclass.py......# |-- Utilities to operate on pytree dataclasses.\n |-- tracing.py...............# |-- Utilities to use when JAX is tracing functions.\n `-- wrappers.py..............# `-- Utilities to wrap objects for specific use cases on pytree dataclass attributes.\n```\n\n</details>\n\n## Credits\n\nThe RBDAs are based on the theory of the [Rigid Body Dynamics Algorithms][RBDA]\nbook by Roy Featherstone.\nThe algorithms and some simulation features were inspired by its accompanying [code][spatial_v2].\n\n[RBDA]: https://link.springer.com/book/10.1007/978-1-4899-7560-7\n[spatial_v2]: http://royfeatherstone.org/spatial/index.html#spatial-software\n\nThe development of JaxSim started in late 2021, inspired by early versions of [`google/brax`][brax].\nAt that time, Brax was implemented in maximal coordinates, and we wanted a physics engine in reduced coordinates.\nWe are grateful to the Brax team for their work and showing the potential of [JAX][jax] in this field.\n\nBrax v2 was later implemented reduced coordinates, following an approach comparable to JaxSim.\nThe development then shifted to [MJX][mjx], which today provides a JAX-based implementation of the Mujoco APIs.\n\nThe main differences between MJX/Brax and JaxSim are as follows:\n\n- JaxSim supports out-of-the-box all SDF models with [Pose Frame Semantics][PFS].\n- JaxSim only supports collisions between points rigidly attached to bodies and a compliant ground surface.\n Our contact model requires careful tuning of its spring-damper parameters, but being an instantaneous\n function of the state $(\\mathbf{q}, \\boldsymbol{\\nu})$, it doesn't require running any optimization algorithm\n when stepping the simulation forward.\n- JaxSim mitigates the stiffness of the contact-aware system dynamics by providing variable-step integrators.\n\n[brax]: https://github.com/google/brax\n[mjx]: https://mujoco.readthedocs.io/en/3.0.0/mjx.html\n[PFS]: http://sdformat.org/tutorials?tut=pose_frame_semantics\n\n## Contributing\n\nWe welcome contributions from the community.\nPlease read the [contributing guide](./CONTRIBUTING.md) to get started.\n\n## Citing\n\n```bibtex\n@software{ferigo_jaxsim_2022,\n author = {Diego Ferigo and Filippo Luca Ferretti and Silvio Traversaro and Daniele Pucci},\n title = {{JaxSim}: A Differentiable Physics Engine and Multibody Dynamics Library for Control and Robot Learning},\n url = {http://github.com/ami-iit/jaxsim},\n year = {2022},\n}\n```\n\nTheoretical aspects of JaxSim are based on Chapters 7 and 8 of the following Ph.D. thesis:\n\n```bibtex\n@phdthesis{ferigo_phd_thesis_2022,\n title = {Simulation Architectures for Reinforcement Learning applied to Robotics},\n author = {Diego Ferigo},\n school = {University of Manchester},\n type = {PhD Thesis},\n month = {July},\n year = {2022},\n}\n```\n\n## People\n\n| Authors | Maintainers |\n|:------:|:-----------:|\n| [<img src=\"https://avatars.githubusercontent.com/u/469199?v=4\" width=\"40\">][df] [<img src=\"https://avatars.githubusercontent.com/u/102977828?v=4\" width=\"40\">][ff] | [<img src=\"https://avatars.githubusercontent.com/u/102977828?v=4\" width=\"40\">][ff] [<img src=\"https://avatars.githubusercontent.com/u/57228872?v=4\" width=\"40\">][ac] |\n\n[df]: https://github.com/diegoferigo\n[ff]: https://github.com/flferretti\n[ac]: https://github.com/xela-95\n\n## License\n\n[BSD3](https://choosealicense.com/licenses/bsd-3-clause/)\n",
"bugtrack_url": null,
"license": "BSD 3-Clause License Copyright (c) 2022, Artificial and Mechanical Intelligence All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. 3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS \"AS IS\" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ",
"summary": "A differentiable physics engine and multibody dynamics library for control and robot learning.",
"version": "0.5.0",
"project_urls": {
"Changelog": "https://github.com/ami-iit/jaxsim/releases",
"Documentation": "https://jaxsim.readthedocs.io",
"Source": "https://github.com/ami-iit/jaxsim",
"Tracker": "https://github.com/ami-iit/jaxsim/issues"
},
"split_keywords": [
"physics",
" physics engine",
" jax",
" rigid body dynamics",
" featherstone",
" reinforcement learning",
" robot",
" robotics",
" sdf",
" urdf"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "05dc6763584ea5834a0ba9d28da0c6e778ce226e283173b03a5c4c0587126a80",
"md5": "9dac1cc2b99ad8e7e39568258f68a7d9",
"sha256": "1387e2a5ae941641aed63d9ee52fb4fe0a600585391b48bd2bf63382ab7e94c6"
},
"downloads": -1,
"filename": "jaxsim-0.5.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "9dac1cc2b99ad8e7e39568258f68a7d9",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.10",
"size": 172767,
"upload_time": "2024-11-15T17:47:55",
"upload_time_iso_8601": "2024-11-15T17:47:55.301731Z",
"url": "https://files.pythonhosted.org/packages/05/dc/6763584ea5834a0ba9d28da0c6e778ce226e283173b03a5c4c0587126a80/jaxsim-0.5.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "e001d88d7844ac5f5fb93510ce4de285b18fbe567c32a058ee32d3ebc8cc06e6",
"md5": "771649d6a6c0ff843c653712c8eae1e1",
"sha256": "656ccf32d4c50cc16754364f58600b3736e838fbee2cd68c26e410954ddab775"
},
"downloads": -1,
"filename": "jaxsim-0.5.0.tar.gz",
"has_sig": false,
"md5_digest": "771649d6a6c0ff843c653712c8eae1e1",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.10",
"size": 195915,
"upload_time": "2024-11-15T17:47:57",
"upload_time_iso_8601": "2024-11-15T17:47:57.372780Z",
"url": "https://files.pythonhosted.org/packages/e0/01/d88d7844ac5f5fb93510ce4de285b18fbe567c32a058ee32d3ebc8cc06e6/jaxsim-0.5.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-11-15 17:47:57",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "ami-iit",
"github_project": "jaxsim",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "jaxsim"
}
