chungmin99/pyroki
chungmin99/pyroki has 1,825 GitHub stars and 209 forks. GitHot measured 0 stars in 24 hours, +4 in 7 days, and +9 in 15 days. Last checked 2026-11-20 20:48 UTC. Last push 2026-04-01.
측정된 스타 증가
- in 24h
- 0
- in 7 days
- +4
- in 15 days
- +9
- in 1 month
- 기록 수집 중
서로 겹치는 기간의 순증가량이며 일별 합계가 아닙니다. 측정값이 없으면 비워 둡니다. 스타는 관심도를 나타내며 품질을 보장하지 않습니다.
소개
A Modular Toolkit for Robot Kinematic Optimization
개발 도구README.md
OriginalPyRoki: Python Robot Kinematics Library
PyRoki is a modular, extensible, and cross-platform toolkit for kinematic optimization, all in Python.
Core features include:
- Differentiable robot forward kinematics model from a URDF.
- Automatic generation of robot collision primitives (e.g., capsules).
- Differentiable collision bodies with numpy broadcasting logic.
- Common cost implementations (e.g., end effector pose, self/world-collision, manipulability).
- Arbitrary costs, autodiff or analytical Jacobians.
- Integration with a Levenberg-Marquardt Solver that supports optimization on manifolds (e.g., lie groups) and hard constraints via an Augmented Lagrangian solver.
- Cross-platform support (CPU, GPU, TPU) via JAX.
Please refer to the documentation for more details, features, and usage examples.
Installation
You can install pyroki with pip, on Python 3.10+:
git clone https://github.com/chungmin99/pyroki.git
cd pyroki
pip install -e .
Status
May 6, 2025: Initial release
We are preparing and will release by May 16, 2025:
- Examples + documentation for hand / humanoid motion retargeting
- Documentation for using manually defined Jacobians
- Support with Python 3.10+
Limitations
- Static shapes & JIT overhead: JAX JIT compilation is triggered on first run and when input shapes change (e.g., number of targets, obstacles). Arrays can be pre-padded to vectorize over inputs with different shapes.
- No sampling-based planners: We don't include sampling-based planners (e.g., graphs, trees).
- Collision performance: Speed and accuracy comparisons against other robot toolkits such as CuRobo have not been extensively performed, and is likely slower than other toolkits for collision-heavy scenarios.
The following are current implementation limitations that could potentially be addressed in future versions:
- Joint types: We only support revolute, continuous, prismatic, and fixed joints. Other URDF joint types are treated as fixed joints.
- Collision geometry: We are limited to sphere, capsule, halfspace, and heightmap geometries. Mesh collision is approximated as capsules.
- Kinematic structures: We only support kinematic trees; no closed-loop mechanisms or parallel manipulators.
Citation
This codebase is released with the following preprint.
| Chung Min Kim*, Brent Yi*, Hongsuk Choi, Yi Ma, Ken Goldberg, Angjoo Kanazawa. PyRoki: A Modular Toolkit for Robot Kinematic Optimization arXiV, 2025. |
*Equal Contribution, UC Berkeley.
Please cite PyRoki if you find this work useful for your research:
@inproceedings{kim2025pyroki,
title={PyRoki: A Modular Toolkit for Robot Kinematic Optimization},
author={Kim*, Chung Min and Yi*, Brent and Choi, Hongsuk and Ma, Yi and Goldberg, Ken and Kanazawa, Angjoo},
booktitle={2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
year={2025},
url={https://arxiv.org/abs/2505.03728},
}
Thanks!