QMCube - A Tactile Cube to Explore Hand Interaction Forces in Human Manipulation

Humans manipulate objects by using sophisticated patterns of finger coordination that are not yet fully documented or understood; objects instrumented with force and tactile sensors have been used in the literature to study the interaction forces applied by the hands, but typically cannot detect bot...

Full description

Saved in:
Bibliographic Details
Published in2023 IEEE International Conference on Development and Learning (ICDL) pp. 37 - 42
Main Authors Murtaza, Zain, Bonzini, Aramis Augusto, Althoefer, Kaspar, Jamone, Lorenzo
Format Conference Proceeding
LanguageEnglish
Published IEEE 09.11.2023
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:Humans manipulate objects by using sophisticated patterns of finger coordination that are not yet fully documented or understood; objects instrumented with force and tactile sensors have been used in the literature to study the interaction forces applied by the hands, but typically cannot detect both the location and the applied force of each finger contact. Access to such information may reveal specific behaviors that have been so far overlooked, and contribute to fully characterizing them. In this paper, we present QMCube: an instrumented cube covered with an unprecedented total of 96 3-axis force sensors (16 for each face, in a 4x4 arrangement). With a modular and compact design, QMCube contains several interchangeable parts that permit modifying the overall weight of the device to test different human manipulation strategies. We report characterization experiments that show that QMCube correctly measures both normal and shear forces on multiple contact locations on its surface, and allows to estimate the contribution of individual fingers during manipulation tasks, such as object picking under different weight conditions. We envisage QMCube to be used in several possible applications, including human studies (e.g. object manipulation and handover), monitoring (e.g. grip strength decay), robot benchmarking (e.g. grasping and manipulation), and human-robot interaction.
DOI:10.1109/ICDL55364.2023.10364422