Talking Robotics

Talking Robotics

Tobias Löw

Date:

Speaker

Tobias Löw is a Postdoctoral Scholar in the Personal Robotics Lab at the University of Washington, Seattle. He received his M.Sc. from ETH Zürich in 2020 and his Ph.D. from EPFL in 2025. His research employs geometric algebra, Lie theory, and differential geometry to develop rich geometric representations for robot learning and control, with a central focus on dexterous manipulation. His current work explores how to combine object-centric action representations with inherent symmetries from multiple kinematic chains to bridge the gap between abstract mathematical structures and reliable real-world robot performance.
Speaker Links: Website · Email: tloew@cs.washington.edu

Abstract

Many tasks in human environments require collaborative behavior between multiple kinematic chains, either to provide additional support for carrying bulky objects or to enable the dexterity required for in-hand manipulation. Coordinating these high-degree-of-freedom systems is notoriously difficult. This work presents the theoretical foundations for cooperative task spaces of multi-arm robotic systems based on geometric primitives defined using conformal geometric algebra. Using similarity transformations of these primitives, we derive an abstraction that enables representing complex multi-arm systems analogously to single-arm systems. We derive analytic and geometric Jacobian matrices and demonstrate integration into classical operational space control. Experiments on bimanual manipulators, humanoids, and multi-fingered hands validate the approach in optimal control and teleoperation and show how the geometric primitives naturally encode nullspace structures for secondary control objectives.

Papers covered

Main paper
Cooperative task spaces for multi-arm manipulation control based on similarity transformations. Tobias Löw, Cem Bilaloglu, and Sylvain Calinon. International Journal of Robotics Research, 2026. DOI: 10.1177/02783649261430763 · arXiv

Related papers