Automatic Differentiation for rigid-body-dynamics AlgorithMs
Project description
adam
Automatic Differentiation for rigid-body-dynamics AlgorithMs
adam implements a collection of algorithms for calculating rigid-body dynamics for floating-base robots, in mixed and body fixed representations (see Traversaro's A Unified View of the Equations of Motion used for Control Design of Humanoid Robots) using:
adam employs the automatic differentiation capabilities of these frameworks to compute, if needed, gradients, Jacobian, Hessians of rigid-body dynamics quantities. This approach enables the design of optimal control and reinforcement learning strategies in robotics.
adam is based on Roy Featherstone's Rigid Body Dynamics Algorithms.
Table of contents
🐍 Dependencies
Other requisites are:
urdf_parser_py
jax
casadi
pytorch
numpy
jax2torch
They will be installed in the installation step!
💾 Installation
The installation can be done either using the Python provided by apt (on Debian-based distros) or via conda (on Linux and macOS).
🐍 Installation with pip
Install python3
, if not installed (in Ubuntu 20.04):
sudo apt install python3.8
Create a virtual environment, if you prefer. For example:
pip install virtualenv
python3 -m venv your_virtual_env
source your_virtual_env/bin/activate
Inside the virtual environment, install the library from pip:
-
Install Jax interface:
pip install adam-robotics[jax]
-
Install CasADi interface:
pip install adam-robotics[casadi]
-
Install PyTorch interface:
pip install adam-robotics[pytorch]
-
Install ALL interfaces:
pip install adam-robotics[all]
If you want the last version:
pip install adam-robotics[selected-interface]@git+https://github.com/ami-iit/ADAM
or clone the repo and install:
git clone https://github.com/ami-iit/adam.git
cd adam
pip install .[selected-interface]
📦 Installation with conda
Installation from conda-forge package
mamba create -n adamenv -c conda-forge adam-robotics
If you want to use jax
or pytorch
, just install the corresponding package as well.
[!NOTE] Check also the conda JAX installation guide here
🔨 Installation from repo
Install in a conda environment the required dependencies:
-
Jax interface dependencies:
mamba create -n adamenv -c conda-forge jax numpy lxml prettytable matplotlib urdfdom-py
-
CasADi interface dependencies:
mamba create -n adamenv -c conda-forge casadi numpy lxml prettytable matplotlib urdfdom-py
-
PyTorch interface dependencies:
mamba create -n adamenv -c conda-forge pytorch numpy lxml prettytable matplotlib urdfdom-py jax2torch
-
ALL interfaces dependencies:
mamba create -n adamenv -c conda-forge jax casadi pytorch numpy lxml prettytable matplotlib urdfdom-py jax2torch
Activate the environment, clone the repo and install the library:
mamba activate adamenv
git clone https://github.com/ami-iit/ADAM.git
cd adam
pip install --no-deps .
🚀 Usage
The following are small snippets of the use of adam. More examples are arriving!
Have also a look at the tests
folder.
Jax interface
[!NOTE] Check also the Jax installation guide here
import adam
from adam.jax import KinDynComputations
import icub_models
import numpy as np
import jax.numpy as jnp
from jax import jit, vmap
# if you want to icub-models https://github.com/robotology/icub-models to retrieve the urdf
model_path = icub_models.get_model_file("iCubGazeboV2_5")
# The joint list
joints_name_list = [
'torso_pitch', 'torso_roll', 'torso_yaw', 'l_shoulder_pitch',
'l_shoulder_roll', 'l_shoulder_yaw', 'l_elbow', 'r_shoulder_pitch',
'r_shoulder_roll', 'r_shoulder_yaw', 'r_elbow', 'l_hip_pitch', 'l_hip_roll',
'l_hip_yaw', 'l_knee', 'l_ankle_pitch', 'l_ankle_roll', 'r_hip_pitch',
'r_hip_roll', 'r_hip_yaw', 'r_knee', 'r_ankle_pitch', 'r_ankle_roll'
]
kinDyn = KinDynComputations(model_path, joints_name_list)
# choose the representation, if you want to use the body fixed representation
kinDyn.set_frame_velocity_representation(adam.Representations.BODY_FIXED_REPRESENTATION)
# or, if you want to use the mixed representation (that is the default)
kinDyn.set_frame_velocity_representation(adam.Representations.MIXED_REPRESENTATION)
w_H_b = np.eye(4)
joints = np.ones(len(joints_name_list))
M = kinDyn.mass_matrix(w_H_b, joints)
print(M)
w_H_f = kinDyn.forward_kinematics('frame_name', w_H_b, joints)
# IMPORTANT! The Jax Interface function execution can be slow! We suggest to jit them.
# For example:
def frame_forward_kinematics(w_H_b, joints):
# This is needed since str is not a valid JAX type
return kinDyn.forward_kinematics('frame_name', w_H_b, joints)
jitted_frame_fk = jit(frame_forward_kinematics)
w_H_f = jitted_frame_fk(w_H_b, joints)
# In the same way, the functions can be also vmapped
vmapped_frame_fk = vmap(frame_forward_kinematics, in_axes=(0, 0))
# which can be also jitted
jitted_vmapped_frame_fk = jit(vmapped_frame_fk)
# and called on a batch of data
joints_batch = jnp.tile(joints, (1024, 1))
w_H_b_batch = jnp.tile(w_H_b, (1024, 1, 1))
w_H_f_batch = jitted_vmapped_frame_fk(w_H_b_batch, joints_batch)
[!NOTE] The first call of the jitted function can be slow, since JAX needs to compile the function. Then it will be faster!
CasADi interface
import adam
from adam.casadi import KinDynComputations
import icub_models
import numpy as np
# if you want to icub-models https://github.com/robotology/icub-models to retrieve the urdf
model_path = icub_models.get_model_file("iCubGazeboV2_5")
# The joint list
joints_name_list = [
'torso_pitch', 'torso_roll', 'torso_yaw', 'l_shoulder_pitch',
'l_shoulder_roll', 'l_shoulder_yaw', 'l_elbow', 'r_shoulder_pitch',
'r_shoulder_roll', 'r_shoulder_yaw', 'r_elbow', 'l_hip_pitch', 'l_hip_roll',
'l_hip_yaw', 'l_knee', 'l_ankle_pitch', 'l_ankle_roll', 'r_hip_pitch',
'r_hip_roll', 'r_hip_yaw', 'r_knee', 'r_ankle_pitch', 'r_ankle_roll'
]
kinDyn = KinDynComputations(model_path, joints_name_list)
# choose the representation you want to use the body fixed representation
kinDyn.set_frame_velocity_representation(adam.Representations.BODY_FIXED_REPRESENTATION)
# or, if you want to use the mixed representation (that is the default)
kinDyn.set_frame_velocity_representation(adam.Representations.MIXED_REPRESENTATION)
w_H_b = np.eye(4)
joints = np.ones(len(joints_name_list))
M = kinDyn.mass_matrix_fun()
print(M(w_H_b, joints))
# If you want to use the symbolic version
w_H_b = cs.SX.eye(4)
joints = cs.SX.sym('joints', len(joints_name_list))
M = kinDyn.mass_matrix_fun()
print(M(w_H_b, joints))
# This is usable also with casadi.MX
w_H_b = cs.MX.eye(4)
joints = cs.MX.sym('joints', len(joints_name_list))
M = kinDyn.mass_matrix_fun()
print(M(w_H_b, joints))
PyTorch interface
import adam
from adam.pytorch import KinDynComputations
import icub_models
import numpy as np
# if you want to icub-models https://github.com/robotology/icub-models to retrieve the urdf
model_path = icub_models.get_model_file("iCubGazeboV2_5")
# The joint list
joints_name_list = [
'torso_pitch', 'torso_roll', 'torso_yaw', 'l_shoulder_pitch',
'l_shoulder_roll', 'l_shoulder_yaw', 'l_elbow', 'r_shoulder_pitch',
'r_shoulder_roll', 'r_shoulder_yaw', 'r_elbow', 'l_hip_pitch', 'l_hip_roll',
'l_hip_yaw', 'l_knee', 'l_ankle_pitch', 'l_ankle_roll', 'r_hip_pitch',
'r_hip_roll', 'r_hip_yaw', 'r_knee', 'r_ankle_pitch', 'r_ankle_roll'
]
kinDyn = KinDynComputations(model_path, joints_name_list)
# choose the representation you want to use the body fixed representation
kinDyn.set_frame_velocity_representation(adam.Representations.BODY_FIXED_REPRESENTATION)
# or, if you want to use the mixed representation (that is the default)
kinDyn.set_frame_velocity_representation(adam.Representations.MIXED_REPRESENTATION)
w_H_b = np.eye(4)
joints = np.ones(len(joints_name_list))
M = kinDyn.mass_matrix(w_H_b, joints)
print(M)
PyTorch Batched interface
[!NOTE] When using this interface, note that the first call of the jitted function can be slow, since JAX needs to compile the function. Then it will be faster!
import adam
from adam.pytorch import KinDynComputationsBatch
import icub_models
# if you want to icub-models
model_path = icub_models.get_model_file("iCubGazeboV2_5")
# The joint list
joints_name_list = [
'torso_pitch', 'torso_roll', 'torso_yaw', 'l_shoulder_pitch',
'l_shoulder_roll', 'l_shoulder_yaw', 'l_elbow', 'r_shoulder_pitch',
'r_shoulder_roll', 'r_shoulder_yaw', 'r_elbow', 'l_hip_pitch', 'l_hip_roll',
'l_hip_yaw', 'l_knee', 'l_ankle_pitch', 'l_ankle_roll', 'r_hip_pitch',
'r_hip_roll', 'r_hip_yaw', 'r_knee', 'r_ankle_pitch', 'r_ankle_roll'
]
kinDyn = KinDynComputationsBatch(model_path, joints_name_list)
# choose the representation you want to use the body fixed representation
kinDyn.set_frame_velocity_representation(adam.Representations.BODY_FIXED_REPRESENTATION)
# or, if you want to use the mixed representation (that is the default)
kinDyn.set_frame_velocity_representation(adam.Representations.MIXED_REPRESENTATION)
w_H_b = np.eye(4)
joints = np.ones(len(joints_name_list))
num_samples = 1024
w_H_b_batch = torch.tensor(np.tile(w_H_b, (num_samples, 1, 1)), dtype=torch.float32)
joints_batch = torch.tensor(np.tile(joints, (num_samples, 1)), dtype=torch.float32)
M = kinDyn.mass_matrix(w_H_b_batch, joints_batch)
w_H_f = kinDyn.forward_kinematics('frame_name', w_H_b_batch, joints_batch)
🦸♂️ Contributing
adam is an open-source project. Contributions are very welcome!
Open an issue with your feature request or if you spot a bug. Then, you can also proceed with a Pull-requests! :rocket:
[!WARNING] REPOSITORY UNDER DEVELOPMENT! We cannot guarantee stable API
Todo
- Center of Mass position
- Jacobians
- Forward kinematics
- Mass Matrix via CRBA
- Centroidal Momentum Matrix via CRBA
- Recursive Newton-Euler algorithm (still no acceleration in the algorithm, since it is used only for the computation of the bias force)
- Articulated Body algorithm
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file adam_robotics-0.3.1.tar.gz
.
File metadata
- Download URL: adam_robotics-0.3.1.tar.gz
- Upload date:
- Size: 69.9 kB
- Tags: Source
- Uploaded using Trusted Publishing? Yes
- Uploaded via: twine/5.1.1 CPython/3.12.7
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | d7613346238e80fe2841ebd92142f7a9022a438b5e8c7b93aa31469f35a1797b |
|
MD5 | 838ad2f609f25df6172e044a746a0a14 |
|
BLAKE2b-256 | bcb8ba22ce5a0d5131c47da68404fcd4bf3c988adef668923e1560868f9b1567 |
Provenance
The following attestation bundles were made for adam_robotics-0.3.1.tar.gz
:
Publisher:
pypi-ci.yml
on ami-iit/adam
-
Statement type:
https://in-toto.io/Statement/v1
- Predicate type:
https://docs.pypi.org/attestations/publish/v1
- Subject name:
adam_robotics-0.3.1.tar.gz
- Subject digest:
d7613346238e80fe2841ebd92142f7a9022a438b5e8c7b93aa31469f35a1797b
- Sigstore transparency entry: 145121283
- Sigstore integration time:
- Predicate type:
File details
Details for the file adam_robotics-0.3.1-py3-none-any.whl
.
File metadata
- Download URL: adam_robotics-0.3.1-py3-none-any.whl
- Upload date:
- Size: 59.2 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? Yes
- Uploaded via: twine/5.1.1 CPython/3.12.7
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | bc1aec918bb2e834199f156f2d5b918372b47c47edce49aff94ac2e395a27f2a |
|
MD5 | 49fa31ddde1603082aa7d4ce9c3a6d2b |
|
BLAKE2b-256 | b7e91c87281b3c189d4ffe30deb41042f7f39acebe10fb278898b215a64e1d30 |
Provenance
The following attestation bundles were made for adam_robotics-0.3.1-py3-none-any.whl
:
Publisher:
pypi-ci.yml
on ami-iit/adam
-
Statement type:
https://in-toto.io/Statement/v1
- Predicate type:
https://docs.pypi.org/attestations/publish/v1
- Subject name:
adam_robotics-0.3.1-py3-none-any.whl
- Subject digest:
bc1aec918bb2e834199f156f2d5b918372b47c47edce49aff94ac2e395a27f2a
- Sigstore transparency entry: 145121285
- Sigstore integration time:
- Predicate type: