jaxsim 0.2.dev423__py3-none-any.whl → 0.2.dev428__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- jaxsim/_version.py +2 -2
- jaxsim-0.2.dev428.dist-info/METADATA +237 -0
- {jaxsim-0.2.dev423.dist-info → jaxsim-0.2.dev428.dist-info}/RECORD +6 -6
- jaxsim-0.2.dev423.dist-info/METADATA +0 -182
- {jaxsim-0.2.dev423.dist-info → jaxsim-0.2.dev428.dist-info}/LICENSE +0 -0
- {jaxsim-0.2.dev423.dist-info → jaxsim-0.2.dev428.dist-info}/WHEEL +0 -0
- {jaxsim-0.2.dev423.dist-info → jaxsim-0.2.dev428.dist-info}/top_level.txt +0 -0
jaxsim/_version.py
CHANGED
@@ -12,5 +12,5 @@ __version__: str
|
|
12
12
|
__version_tuple__: VERSION_TUPLE
|
13
13
|
version_tuple: VERSION_TUPLE
|
14
14
|
|
15
|
-
__version__ = version = '0.2.
|
16
|
-
__version_tuple__ = version_tuple = (0, 2, '
|
15
|
+
__version__ = version = '0.2.dev428'
|
16
|
+
__version_tuple__ = version_tuple = (0, 2, 'dev428')
|
@@ -0,0 +1,237 @@
|
|
1
|
+
Metadata-Version: 2.1
|
2
|
+
Name: jaxsim
|
3
|
+
Version: 0.2.dev428
|
4
|
+
Summary: A differentiable physics engine and multibody dynamics library for control and robot learning.
|
5
|
+
Home-page: https://github.com/ami-iit/jaxsim
|
6
|
+
Author: Diego Ferigo
|
7
|
+
Author-email: diego.ferigo@iit.it
|
8
|
+
License: BSD
|
9
|
+
Project-URL: Changelog, https://github.com/ami-iit/jaxsim/releases
|
10
|
+
Project-URL: Documentation, https://jaxsim.readthedocs.io
|
11
|
+
Project-URL: Source, https://github.com/ami-iit/jaxsim
|
12
|
+
Project-URL: Tracker, https://github.com/ami-iit/jaxsim/issues
|
13
|
+
Keywords: physics,physics engine,jax,rigid body dynamics,featherstone,reinforcement learning,robot,robotics,sdf,urdf
|
14
|
+
Platform: any
|
15
|
+
Classifier: Development Status :: 4 - Beta
|
16
|
+
Classifier: Framework :: Robot Framework
|
17
|
+
Classifier: Intended Audience :: Developers
|
18
|
+
Classifier: Intended Audience :: Science/Research
|
19
|
+
Classifier: License :: OSI Approved :: BSD License
|
20
|
+
Classifier: Operating System :: POSIX :: Linux
|
21
|
+
Classifier: Operating System :: MacOS
|
22
|
+
Classifier: Operating System :: Microsoft
|
23
|
+
Classifier: Programming Language :: Python :: 3.11
|
24
|
+
Classifier: Programming Language :: Python :: 3.12
|
25
|
+
Classifier: Programming Language :: Python :: 3 :: Only
|
26
|
+
Classifier: Programming Language :: Python :: Implementation :: CPython
|
27
|
+
Classifier: Topic :: Games/Entertainment :: Simulation
|
28
|
+
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
|
29
|
+
Classifier: Topic :: Scientific/Engineering :: Physics
|
30
|
+
Classifier: Topic :: Software Development
|
31
|
+
Requires-Python: >=3.11
|
32
|
+
Description-Content-Type: text/markdown
|
33
|
+
License-File: LICENSE
|
34
|
+
Requires-Dist: coloredlogs
|
35
|
+
Requires-Dist: jax >=0.4.13
|
36
|
+
Requires-Dist: jaxlib >=0.4.13
|
37
|
+
Requires-Dist: jaxlie >=1.3.0
|
38
|
+
Requires-Dist: jax-dataclasses >=1.4.0
|
39
|
+
Requires-Dist: pptree
|
40
|
+
Requires-Dist: rod >=0.2.0
|
41
|
+
Requires-Dist: typing-extensions ; python_version < "3.12"
|
42
|
+
Provides-Extra: all
|
43
|
+
Requires-Dist: black[jupyter] ~=24.0 ; extra == 'all'
|
44
|
+
Requires-Dist: isort ; extra == 'all'
|
45
|
+
Requires-Dist: pre-commit ; extra == 'all'
|
46
|
+
Requires-Dist: idyntree ; extra == 'all'
|
47
|
+
Requires-Dist: pytest >=6.0 ; extra == 'all'
|
48
|
+
Requires-Dist: pytest-icdiff ; extra == 'all'
|
49
|
+
Requires-Dist: robot-descriptions ; extra == 'all'
|
50
|
+
Requires-Dist: lxml ; extra == 'all'
|
51
|
+
Requires-Dist: mediapy ; extra == 'all'
|
52
|
+
Requires-Dist: mujoco >=3.0.0 ; extra == 'all'
|
53
|
+
Provides-Extra: style
|
54
|
+
Requires-Dist: black[jupyter] ~=24.0 ; extra == 'style'
|
55
|
+
Requires-Dist: isort ; extra == 'style'
|
56
|
+
Requires-Dist: pre-commit ; extra == 'style'
|
57
|
+
Provides-Extra: testing
|
58
|
+
Requires-Dist: idyntree ; extra == 'testing'
|
59
|
+
Requires-Dist: pytest >=6.0 ; extra == 'testing'
|
60
|
+
Requires-Dist: pytest-icdiff ; extra == 'testing'
|
61
|
+
Requires-Dist: robot-descriptions ; extra == 'testing'
|
62
|
+
Provides-Extra: viz
|
63
|
+
Requires-Dist: lxml ; extra == 'viz'
|
64
|
+
Requires-Dist: mediapy ; extra == 'viz'
|
65
|
+
Requires-Dist: mujoco >=3.0.0 ; extra == 'viz'
|
66
|
+
|
67
|
+
# JaxSim
|
68
|
+
|
69
|
+
JaxSim is a **differentiable physics engine** and **multibody dynamics library** designed for applications in control and robot learning, implemented with JAX.
|
70
|
+
|
71
|
+
Its design facilitates research and accelerates prototyping in the intersection of robotics and artificial intelligence.
|
72
|
+
|
73
|
+
## Features
|
74
|
+
|
75
|
+
- Physics engine in reduced coordinates supporting fixed-base and floating-base robots.
|
76
|
+
- Multibody dynamics library providing all the necessary components for developing model-based control algorithms.
|
77
|
+
- Completely developed in Python with [`google/jax`][jax] following a functional programming paradigm.
|
78
|
+
- Transparent support for running on CPUs, GPUs, and TPUs.
|
79
|
+
- Full support for JIT compilation for increased performance.
|
80
|
+
- Full support for automatic vectorization for massive parallelization of open-loop and closed-loop architectures.
|
81
|
+
- Support for SDF models and, upon conversion with [sdformat][sdformat], URDF models.
|
82
|
+
- Visualization based on the [passive viewer][passive_viewer_mujoco] of Mujoco.
|
83
|
+
|
84
|
+
### JaxSim as a simulator
|
85
|
+
|
86
|
+
- Wide range of fixed-step explicit Runge-Kutta integrators.
|
87
|
+
- Support for variable-step integrators implemented as embedded Runge-Kutta schemes.
|
88
|
+
- Improved stability by optionally integrating the base orientation on the $\text{SO}(3)$ manifold.
|
89
|
+
- Soft contacts model supporting full friction cone and sticking-slipping transition.
|
90
|
+
- Collision detection between points rigidly attached to bodies and uneven ground surfaces.
|
91
|
+
|
92
|
+
### JaxSim as a multibody dynamics library
|
93
|
+
|
94
|
+
- Provides rigid body dynamics algorithms (RBDAs) like RNEA, ABA, CRBA, and Jacobians.
|
95
|
+
- Provides all the quantities included in the Euler-Poincarè formulation of the equations of motion.
|
96
|
+
- Supports body-fixed, inertial-fixed, and mixed [velocity representations][notation].
|
97
|
+
- Exposes all the necessary quantities to develop controllers in centroidal coordinates.
|
98
|
+
|
99
|
+
### JaxSim for robot learning
|
100
|
+
|
101
|
+
- Being developed with JAX, all the RBDAs support automatic differentiation both in forward and reverse modes.
|
102
|
+
- Support for automatically differentiating against kinematics and dynamics parameters.
|
103
|
+
- All fixed-step integrators are forward and reverse differentiable.
|
104
|
+
- All variable-step integrators are forward differentiable.
|
105
|
+
- Ideal for sampling synthetic data for reinforcement learning (RL).
|
106
|
+
- Ideal for designing physics-informed neural networks (PINNs) with loss functions requiring model-based quantities.
|
107
|
+
- Ideal for combining model-based control with learning-based components.
|
108
|
+
|
109
|
+
[jax]: https://github.com/google/jax/
|
110
|
+
[sdformat]: https://github.com/gazebosim/sdformat
|
111
|
+
[notation]: https://research.tue.nl/en/publications/multibody-dynamics-notation-version-2
|
112
|
+
[passive_viewer_mujoco]: https://mujoco.readthedocs.io/en/stable/python.html#passive-viewer
|
113
|
+
|
114
|
+
> [!WARNING]
|
115
|
+
> This project is still experimental, APIs could change between releases without notice.
|
116
|
+
|
117
|
+
> [!NOTE]
|
118
|
+
> JaxSim currently focuses on locomotion applications.
|
119
|
+
> Only contacts between bodies and smooth ground surfaces are supported.
|
120
|
+
|
121
|
+
## Documentation
|
122
|
+
|
123
|
+
The JaxSim API documentation is available at [jaxsim.readthedocs.io][readthedocs].
|
124
|
+
|
125
|
+
[readthedocs]: https://jaxsim.readthedocs.io/
|
126
|
+
|
127
|
+
## Installation
|
128
|
+
|
129
|
+
<details>
|
130
|
+
<summary>With conda</summary>
|
131
|
+
|
132
|
+
You can install the project using [`conda`][conda] as follows:
|
133
|
+
|
134
|
+
```bash
|
135
|
+
conda install jaxsim -c conda-forge
|
136
|
+
```
|
137
|
+
|
138
|
+
You can enforce GPU support, if needed, by also specifying `"jaxlib = * = *cuda*"`.
|
139
|
+
|
140
|
+
</details>
|
141
|
+
|
142
|
+
<details>
|
143
|
+
<summary>With pip</summary>
|
144
|
+
|
145
|
+
You can install the project using [`pypa/pip`][pip], preferably in a [virtual environment][venv], as follows:
|
146
|
+
|
147
|
+
```bash
|
148
|
+
pip install jaxsim
|
149
|
+
```
|
150
|
+
|
151
|
+
Check [`setup.cfg`](setup.cfg) for the complete list of optional dependencies.
|
152
|
+
You can obtain a full installation using `jaxsim[all]`.
|
153
|
+
|
154
|
+
If you need GPU support, follow the official [installation instructions][jax_gpu] of JAX.
|
155
|
+
|
156
|
+
</details>
|
157
|
+
|
158
|
+
<details>
|
159
|
+
<summary>Contributors installation</summary>
|
160
|
+
|
161
|
+
If you want to contribute to the project, we recommend creating the following `jaxsim` conda environment first:
|
162
|
+
|
163
|
+
```bash
|
164
|
+
conda env create -f environment.yml
|
165
|
+
```
|
166
|
+
|
167
|
+
Then, activate the environment and install the project in editable mode:
|
168
|
+
|
169
|
+
```bash
|
170
|
+
conda activate jaxsim
|
171
|
+
pip install --no-deps -e .
|
172
|
+
```
|
173
|
+
|
174
|
+
</details>
|
175
|
+
|
176
|
+
[conda]: https://anaconda.org/
|
177
|
+
[pip]: https://github.com/pypa/pip/
|
178
|
+
[venv]: https://docs.python.org/3/tutorial/venv.html
|
179
|
+
[jax_gpu]: https://github.com/google/jax/#installation
|
180
|
+
|
181
|
+
## Credits
|
182
|
+
|
183
|
+
The RBDAs are based on the theory of the [Rigid Body Dynamics Algorithms][RBDA]
|
184
|
+
book by Roy Featherstone.
|
185
|
+
The algorithms and some simulation features were inspired by its accompanying [code][spatial_v2].
|
186
|
+
|
187
|
+
[RBDA]: https://link.springer.com/book/10.1007/978-1-4899-7560-7
|
188
|
+
[spatial_v2]: http://royfeatherstone.org/spatial/index.html#spatial-software
|
189
|
+
|
190
|
+
The development of JaxSim started in late 2021, inspired by early versions of [`google/brax`][brax].
|
191
|
+
At that time, Brax was implemented in maximal coordinates, and we wanted a physics engine in reduced coordinates.
|
192
|
+
We are grateful to the Brax team for their work and showing the potential of [JAX][jax] in this field.
|
193
|
+
|
194
|
+
Brax v2 was later implemented reduced coordinates, following an approach comparable to JaxSim.
|
195
|
+
The development then shifted to [MJX][mjx], which today provides a JAX-based implementation of the Mujoco APIs.
|
196
|
+
|
197
|
+
The main differences between MJX/Brax and JaxSim are as follows:
|
198
|
+
|
199
|
+
- JaxSim supports out-of-the-box all SDF models with [Pose Frame Semantics][PFS].
|
200
|
+
- JaxSim only supports collisions between points rigidly attached to bodies and a compliant ground surface.
|
201
|
+
Our contact model requires careful tuning of its spring-damper parameters, but being an instantaneous
|
202
|
+
function of the state $(\mathbf{q}, \boldsymbol{\nu})$, it doesn't require running any optimization algorithm
|
203
|
+
when stepping the simulation forward.
|
204
|
+
- JaxSim mitigates the stiffness of the contact-aware system dynamics by providing variable-step integrators.
|
205
|
+
|
206
|
+
[brax]: https://github.com/google/brax
|
207
|
+
[mjx]: https://mujoco.readthedocs.io/en/3.0.0/mjx.html
|
208
|
+
[PFS]: http://sdformat.org/tutorials?tut=pose_frame_semantics
|
209
|
+
|
210
|
+
## Contributing
|
211
|
+
|
212
|
+
We welcome contributions from the community.
|
213
|
+
Please read the [contributing guide](./CONTRIBUTING.md) to get started.
|
214
|
+
|
215
|
+
## Citing
|
216
|
+
|
217
|
+
```bibtex
|
218
|
+
@software{ferigo_jaxsim_2022,
|
219
|
+
author = {Diego Ferigo and Filippo Luca Ferretti and Silvio Traversaro and Daniele Pucci},
|
220
|
+
title = {{JaxSim}: A Differentiable Physics Engine and Multibody Dynamics Library for Control and Robot Learning},
|
221
|
+
url = {http://github.com/ami-iit/jaxsim},
|
222
|
+
year = {2022},
|
223
|
+
}
|
224
|
+
```
|
225
|
+
|
226
|
+
## People
|
227
|
+
|
228
|
+
| Author | Maintainers |
|
229
|
+
|:------:|:-----------:|
|
230
|
+
| [<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/469199?v=4" width="40">][df] |
|
231
|
+
|
232
|
+
[df]: https://github.com/diegoferigo
|
233
|
+
[ff]: https://github.com/flferretti
|
234
|
+
|
235
|
+
## License
|
236
|
+
|
237
|
+
[BSD3](https://choosealicense.com/licenses/bsd-3-clause/)
|
@@ -1,5 +1,5 @@
|
|
1
1
|
jaxsim/__init__.py,sha256=OcrfoYS1DGcmAGqu2AqlCTiUVxcpi-IsVwcr_16x74Q,1789
|
2
|
-
jaxsim/_version.py,sha256=
|
2
|
+
jaxsim/_version.py,sha256=1mo3HxrVkFrROneeiBN6Fz-k5UgdU7xlKMwBqx6ub4Q,423
|
3
3
|
jaxsim/logging.py,sha256=c4zhwBKf9eAYAHVp62kTEllqdsZgh0K-kPKVy8L3elU,1584
|
4
4
|
jaxsim/typing.py,sha256=MeuOCQtLAr-sPkvB_sU8FtwGNRirz1auCwIgRC-QZl8,646
|
5
5
|
jaxsim/api/__init__.py,sha256=fNTCPUeDfOAizRd4RsW3Epv0sLTu0KJGoFRSEsi75VM,162
|
@@ -57,8 +57,8 @@ jaxsim/utils/__init__.py,sha256=tnQq1_CavdfeKaLYt3pmO7Jk4MU2RwwQU_qICkjyoTY,197
|
|
57
57
|
jaxsim/utils/hashless.py,sha256=bFIwKeo9KiWwsY8QM55duEGGQOyyJ4jQyPcuqTLEp5k,297
|
58
58
|
jaxsim/utils/jaxsim_dataclass.py,sha256=T8nZolb563mYJD7YvTmYW-z-hpkEiaK3cDbYHqwtqRc,11347
|
59
59
|
jaxsim/utils/tracing.py,sha256=KDMoyVPlu2NJvFkhtZwq5AkqMMgajt3munvJom-vEjQ,650
|
60
|
-
jaxsim-0.2.
|
61
|
-
jaxsim-0.2.
|
62
|
-
jaxsim-0.2.
|
63
|
-
jaxsim-0.2.
|
64
|
-
jaxsim-0.2.
|
60
|
+
jaxsim-0.2.dev428.dist-info/LICENSE,sha256=eaYdFmdeMbiIoIiPzEK0MjP1S9wtFXjXNR5er49uLR0,1546
|
61
|
+
jaxsim-0.2.dev428.dist-info/METADATA,sha256=iMVNQHSzzrTQenqwQnlkCMI9xl8yxW7wLNoKVp1nfmY,9778
|
62
|
+
jaxsim-0.2.dev428.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
|
63
|
+
jaxsim-0.2.dev428.dist-info/top_level.txt,sha256=LxGMA8FLtXjQ6oI7N5gd_R_oSUHxpXxUEOfT1xS_ni0,7
|
64
|
+
jaxsim-0.2.dev428.dist-info/RECORD,,
|
@@ -1,182 +0,0 @@
|
|
1
|
-
Metadata-Version: 2.1
|
2
|
-
Name: jaxsim
|
3
|
-
Version: 0.2.dev423
|
4
|
-
Summary: A physics engine in reduced coordinates implemented with JAX.
|
5
|
-
Home-page: https://github.com/ami-iit/jaxsim
|
6
|
-
Author: Diego Ferigo
|
7
|
-
Author-email: diego.ferigo@iit.it
|
8
|
-
License: BSD
|
9
|
-
Project-URL: Changelog, https://github.com/ami-iit/jaxsim/releases
|
10
|
-
Project-URL: Source, https://github.com/ami-iit/jaxsim
|
11
|
-
Project-URL: Tracker, https://github.com/ami-iit/jaxsim/issues
|
12
|
-
Keywords: physics,physics engine,jax,rigid body dynamics,featherstone,reinforcement learning,robot,robotics,sdf,urdf
|
13
|
-
Platform: any
|
14
|
-
Classifier: Development Status :: 4 - Beta
|
15
|
-
Classifier: Framework :: Robot Framework
|
16
|
-
Classifier: Intended Audience :: Developers
|
17
|
-
Classifier: Intended Audience :: Science/Research
|
18
|
-
Classifier: License :: OSI Approved :: BSD License
|
19
|
-
Classifier: Operating System :: POSIX :: Linux
|
20
|
-
Classifier: Operating System :: MacOS
|
21
|
-
Classifier: Operating System :: Microsoft
|
22
|
-
Classifier: Programming Language :: Python :: 3.11
|
23
|
-
Classifier: Programming Language :: Python :: 3.12
|
24
|
-
Classifier: Programming Language :: Python :: 3 :: Only
|
25
|
-
Classifier: Programming Language :: Python :: Implementation :: CPython
|
26
|
-
Classifier: Topic :: Games/Entertainment :: Simulation
|
27
|
-
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
|
28
|
-
Classifier: Topic :: Scientific/Engineering :: Physics
|
29
|
-
Classifier: Topic :: Software Development
|
30
|
-
Requires-Python: >=3.11
|
31
|
-
Description-Content-Type: text/markdown
|
32
|
-
License-File: LICENSE
|
33
|
-
Requires-Dist: coloredlogs
|
34
|
-
Requires-Dist: jax >=0.4.13
|
35
|
-
Requires-Dist: jaxlib >=0.4.13
|
36
|
-
Requires-Dist: jaxlie >=1.3.0
|
37
|
-
Requires-Dist: jax-dataclasses >=1.4.0
|
38
|
-
Requires-Dist: pptree
|
39
|
-
Requires-Dist: rod >=0.2.0
|
40
|
-
Requires-Dist: typing-extensions ; python_version < "3.12"
|
41
|
-
Provides-Extra: all
|
42
|
-
Requires-Dist: black[jupyter] ~=24.0 ; extra == 'all'
|
43
|
-
Requires-Dist: isort ; extra == 'all'
|
44
|
-
Requires-Dist: pre-commit ; extra == 'all'
|
45
|
-
Requires-Dist: idyntree ; extra == 'all'
|
46
|
-
Requires-Dist: pytest >=6.0 ; extra == 'all'
|
47
|
-
Requires-Dist: pytest-icdiff ; extra == 'all'
|
48
|
-
Requires-Dist: robot-descriptions ; extra == 'all'
|
49
|
-
Requires-Dist: lxml ; extra == 'all'
|
50
|
-
Requires-Dist: mediapy ; extra == 'all'
|
51
|
-
Requires-Dist: mujoco >=3.0.0 ; extra == 'all'
|
52
|
-
Provides-Extra: style
|
53
|
-
Requires-Dist: black[jupyter] ~=24.0 ; extra == 'style'
|
54
|
-
Requires-Dist: isort ; extra == 'style'
|
55
|
-
Requires-Dist: pre-commit ; extra == 'style'
|
56
|
-
Provides-Extra: testing
|
57
|
-
Requires-Dist: idyntree ; extra == 'testing'
|
58
|
-
Requires-Dist: pytest >=6.0 ; extra == 'testing'
|
59
|
-
Requires-Dist: pytest-icdiff ; extra == 'testing'
|
60
|
-
Requires-Dist: robot-descriptions ; extra == 'testing'
|
61
|
-
Provides-Extra: viz
|
62
|
-
Requires-Dist: lxml ; extra == 'viz'
|
63
|
-
Requires-Dist: mediapy ; extra == 'viz'
|
64
|
-
Requires-Dist: mujoco >=3.0.0 ; extra == 'viz'
|
65
|
-
|
66
|
-
# JAXsim
|
67
|
-
|
68
|
-
**A scalable physics engine and multibody dynamics library implemented with JAX. With JIT batteries 🔋**
|
69
|
-
|
70
|
-
> [!WARNING]
|
71
|
-
> This project is still experimental, APIs could change without notice.
|
72
|
-
|
73
|
-
> [!NOTE]
|
74
|
-
> This simulator currently focuses on locomotion applications. Only contacts with ground are supported.
|
75
|
-
|
76
|
-
## Features
|
77
|
-
|
78
|
-
- Physics engine in reduced coordinates implemented with [JAX][jax] in Python.
|
79
|
-
- JIT compilation of Python code for increased performance.
|
80
|
-
- Transparent support to execute logic on CPUs, GPUs, and TPUs.
|
81
|
-
- Parallel multi-body simulations on hardware accelerators for significantly increased throughput.
|
82
|
-
- Support for SDF models (and, upon conversion, URDF models).
|
83
|
-
- Collision detection between bodies and uneven ground surface.
|
84
|
-
- Soft contacts model supporting full friction cone and sticking / slipping transition.
|
85
|
-
- Complete support for inertial properties of rigid bodies.
|
86
|
-
- Revolute, prismatic, and fixed joints support.
|
87
|
-
- Integrators: forward Euler, semi-implicit Euler, Runge-Kutta 4.
|
88
|
-
- High-level classes for object-oriented programming.
|
89
|
-
- High-level classes to compute multi-body dynamics quantities from the simulation state.
|
90
|
-
- High-level classes wrapping the low-level functional RBDAs with support of [multiple velocities representations][notation].
|
91
|
-
- Default validation of JAX pytrees to prevent JIT re-compilations.
|
92
|
-
- Preliminary support for automatic differentiation of RBDAs.
|
93
|
-
|
94
|
-
[jax]: https://github.com/google/jax/
|
95
|
-
[notation]: https://research.tue.nl/en/publications/multibody-dynamics-notation-version-2
|
96
|
-
|
97
|
-
## Documentation
|
98
|
-
|
99
|
-
The JAXsim API documentation is available at [jaxsim.readthedocs.io](https://jaxsim.readthedocs.io/).
|
100
|
-
|
101
|
-
## Installation
|
102
|
-
|
103
|
-
You can install the project using [`conda`][conda]:
|
104
|
-
|
105
|
-
```bash
|
106
|
-
conda install jaxsim -c conda-forge
|
107
|
-
```
|
108
|
-
|
109
|
-
Alternatively, you can use [`pypa/pip`][pip], preferably in a [virtual environment][venv]:
|
110
|
-
|
111
|
-
```bash
|
112
|
-
pip install jaxsim
|
113
|
-
```
|
114
|
-
|
115
|
-
Check [`setup.cfg`](setup.cfg) for the complete list of optional dependencies.
|
116
|
-
Install all of them with `jaxsim[all]`.
|
117
|
-
|
118
|
-
**Note:** For GPU support, follow the official [installation instructions][jax_gpu] of JAX.
|
119
|
-
|
120
|
-
[conda]: https://anaconda.org/
|
121
|
-
[pip]: https://github.com/pypa/pip/
|
122
|
-
[venv]: https://docs.python.org/3/tutorial/venv.html
|
123
|
-
[jax_gpu]: https://github.com/google/jax/#installation
|
124
|
-
|
125
|
-
## Quickstart
|
126
|
-
|
127
|
-
Explore and learn how to use the library through practical demonstrations available in the [examples](./examples) folder.
|
128
|
-
|
129
|
-
## Credits
|
130
|
-
|
131
|
-
The physics module of JAXsim is based on the theory of the [Rigid Body Dynamics Algorithms][RBDA]
|
132
|
-
book by Roy Featherstone.
|
133
|
-
We structured part of our logic following its accompanying [code][spatial_v2].
|
134
|
-
The physics engine is developed entirely in Python using [JAX][jax].
|
135
|
-
|
136
|
-
[RBDA]: https://link.springer.com/book/10.1007/978-1-4899-7560-7
|
137
|
-
[spatial_v2]: http://royfeatherstone.org/spatial/index.html#spatial-software
|
138
|
-
|
139
|
-
The inspiration for developing JAXsim originally stemmed from early versions of [`google/brax`][brax].
|
140
|
-
Here below we summarize the differences between the projects:
|
141
|
-
|
142
|
-
- JAXsim simulates multibody dynamics in reduced coordinates, while brax v1 uses maximal coordinates.
|
143
|
-
- The new v2 APIs of brax (and the new [MJX][mjx]) were then implemented in reduced coordinates, following an approach comparable to JAXsim, with major differences in contact handling.
|
144
|
-
- The rigid-body algorithms used in JAXsim allow to efficiently compute quantities based on the Euler-Poincarè
|
145
|
-
formulation of the equations of motion, necessary for model-based robotics research.
|
146
|
-
- JAXsim supports SDF (and, indirectly, URDF) models, assuming the model is described with the
|
147
|
-
recent [Pose Frame Semantics][PFS].
|
148
|
-
- Contrarily to brax, JAXsim only supports collision detection between bodies and a compliant ground surface.
|
149
|
-
- The RBDAs of JAXsim support automatic differentiation, but this functionality has not been thoroughly tested.
|
150
|
-
|
151
|
-
[brax]: https://github.com/google/brax
|
152
|
-
[mjx]: https://mujoco.readthedocs.io/en/3.0.0/mjx.html
|
153
|
-
[PFS]: http://sdformat.org/tutorials?tut=pose_frame_semantics
|
154
|
-
|
155
|
-
## Contributing
|
156
|
-
|
157
|
-
We welcome contributions from the community.
|
158
|
-
Please read the [contributing guide](./CONTRIBUTING.md) to get started.
|
159
|
-
|
160
|
-
## Citing
|
161
|
-
|
162
|
-
```bibtex
|
163
|
-
@software{ferigo_jaxsim_2022,
|
164
|
-
author = {Diego Ferigo and Silvio Traversaro and Daniele Pucci},
|
165
|
-
title = {{JAXsim}: A Physics Engine in Reduced Coordinates and Multibody Dynamics Library for Control and Robot Learning},
|
166
|
-
url = {http://github.com/ami-iit/jaxsim},
|
167
|
-
year = {2022},
|
168
|
-
}
|
169
|
-
```
|
170
|
-
|
171
|
-
## People
|
172
|
-
|
173
|
-
| Author | Maintainers |
|
174
|
-
|:------:|:-----------:|
|
175
|
-
| [<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/469199?v=4" width="40">][df] |
|
176
|
-
|
177
|
-
[df]: https://github.com/diegoferigo
|
178
|
-
[ff]: https://github.com/flferretti
|
179
|
-
|
180
|
-
## License
|
181
|
-
|
182
|
-
[BSD3](https://choosealicense.com/licenses/bsd-3-clause/)
|
File without changes
|
File without changes
|
File without changes
|