PyRoki: A Modular Toolkit for Robot Kinematic Optimization

TL;DR

PyRoki is a versatile, cross-platform toolkit for robot kinematic optimization that enables efficient and customizable motion planning, outperforming existing GPU-based solutions in speed and accuracy.

Contribution

PyRoki introduces a modular, extensible framework for kinematic optimization that works across CPU, GPU, and TPU, with demonstrated advantages in speed and error reduction.

Findings

PyRoki is 1.4-1.7x faster than cuRobo.

PyRoki achieves lower optimization errors.

PyRoki's modular design facilitates diverse motion planning applications.

Abstract

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.