PyRoki: A Modular Toolkit for Robot Kinematic Optimization

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Kim, Chung Min, Yi, Brent, Choi, Hongsuk, Ma, Yi, Goldberg, Ken, Kanazawa, Angjoo
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916722830737408
author Kim, Chung Min
Yi, Brent
Choi, Hongsuk
Ma, Yi
Goldberg, Ken
Kanazawa, Angjoo
author_facet Kim, Chung Min
Yi, Brent
Choi, Hongsuk
Ma, Yi
Goldberg, Ken
Kanazawa, Angjoo
contents Robot motion can have many goals. Depending on the task, we might optimize for pose error, speed, collision, or similarity to a human demonstration. Motivated by this, we present PyRoki: a modular, extensible, and cross-platform toolkit for solving kinematic optimization problems. PyRoki couples an interface for specifying kinematic variables and costs with an efficient nonlinear least squares optimizer. Unlike existing tools, it is also cross-platform: optimization runs natively on CPU, GPU, and TPU. In this paper, we present (i) the design and implementation of PyRoki, (ii) motion retargeting and planning case studies that highlight the advantages of PyRoki's modularity, and (iii) optimization benchmarking, where PyRoki can be 1.4-1.7x faster and converges to lower errors than cuRobo, an existing GPU-accelerated inverse kinematics library.
format Preprint
id arxiv_https___arxiv_org_abs_2505_03728
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PyRoki: A Modular Toolkit for Robot Kinematic Optimization
Kim, Chung Min
Yi, Brent
Choi, Hongsuk
Ma, Yi
Goldberg, Ken
Kanazawa, Angjoo
Robotics
Robot motion can have many goals. Depending on the task, we might optimize for pose error, speed, collision, or similarity to a human demonstration. Motivated by this, we present PyRoki: a modular, extensible, and cross-platform toolkit for solving kinematic optimization problems. PyRoki couples an interface for specifying kinematic variables and costs with an efficient nonlinear least squares optimizer. Unlike existing tools, it is also cross-platform: optimization runs natively on CPU, GPU, and TPU. In this paper, we present (i) the design and implementation of PyRoki, (ii) motion retargeting and planning case studies that highlight the advantages of PyRoki's modularity, and (iii) optimization benchmarking, where PyRoki can be 1.4-1.7x faster and converges to lower errors than cuRobo, an existing GPU-accelerated inverse kinematics library.
title PyRoki: A Modular Toolkit for Robot Kinematic Optimization
topic Robotics
url https://arxiv.org/abs/2505.03728