Embodied Geometric Intelligence ( E G i ) for contact in robots

Tafrishi, Seyed Amir ORCID: https://orcid.org/0000-0001-9829-3144 2026. Embodied Geometric Intelligence ( E G i ) for contact in robots. Presented at: EI 2025, Online, 02 - 04 April 2025. IOP Conference Series: Materials Science and Engineering. , vol.1343 (1) IOP Publishing, 10.1088/1757-899x/1343/1/012017

PDF - Published Version Download (392kB)

Abstract

We introduce embodied geometric intelligence (EGi), a paradigm in which a robot’s morphology and passive compliance become active participants in perception, control, and manipulation. Leveraging underactuated or deformable bodies, soft robots exploit geometric cues in contact and environmental compliance to achieve traversal, grasping, and shape morphing without relying solely on onboard sensing or complex controllers. The term “geometry” traces to the Latin geōmetría, “gē” (land) + “metría” (measurement), highlighting knowledge acquired through contact. We formalize this approach using a curvature–compliance model unified with Montana’s contact kinematics, and we analyse the “dead fish in a stream” exemplar as a canonical case of geometry-driven control. Conceptual and mathematical results indicate that geometry-driven behaviors yield robustness and adaptability in dynamic, unstructured environments. This work lays the groundwork for a reimagined embodied intelligence that integrates geometry, physics, and control into a cohesive adaptive body.

Item Type:	Conference or Workshop Item - published (Paper)
Date Type:	Publication
Status:	Published
Schools:	Schools > Engineering
Additional Information:	License information from Publisher: LICENSE 1: URL: https://creativecommons.org/licenses/by/4.0/, Type: cc-by
Publisher:	IOP Publishing
ISSN:	1757-8981
Date of First Compliant Deposit:	8 April 2026
Last Modified:	08 Apr 2026 09:15
URI:	https://orca.cardiff.ac.uk/id/eprint/186260

Actions (repository staff only)

Edit Item