Skip to Main content Skip to Navigation
Theses

Toucher artificiel à base d'un microcapteur d'effort: traitement du signal et des informations associées

Abstract : This document presents the conception, the characterisation and the use of an artificial touch system for a study on tactile texture recognition. The artificial touch system is composed of an artificial finger and an exploration device allowing an automatic translation of the finger on samples. The artificial finger is based on a tri-axis force MEMS sensor covered with a polyurethane artificial skin. Thus, the MEMS measure the three spatial components of the efforts transmitted by the polyurethane coating. To rub the finger against surfaces, two exploration devices have been conceived successively. Thanks to the first one, we rapidly organized some texture recognition experiments, to evaluate the ability of the finger to discriminate texture. The second device, allows a better control of the exploration parameters, like moving speed or forces applied. This device also allows us to characterize the artificial finger during a simple contact task or while rubbing the artificial finger on a surface. A study of that kind allowed us to better understand the physical phenomena taking place during tactile texture exploration. The first experiments on texture exploration showed a good sensitivity of the artificial finger to detect the weave form periodicity of a textile, or to discriminate two different kinds of textures (paper v.s. textile). Next, the artificial finger is evaluated for the discrimination of ten kinds of papers. The signals given by the artificial finger while rubbing a texture are showing some problems of reproducibility. However, associated with several classification algorithms, the artificial finger gives good results for the discrimination of the ten textures, in particular thanks to spectral characteristics extracted from the signals. This experiment was run several times with different protocols, in order to validate the classification results. These results support several hypotheses on the behaviour of the artificial touch system and to some extent on tactile texture recognition.
Complete list of metadatas

Cited literature [80 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00512748
Contributor : Florian de Boissieu <>
Submitted on : Tuesday, August 31, 2010 - 2:54:15 PM
Last modification on : Thursday, November 19, 2020 - 12:59:39 PM
Long-term archiving on: : Wednesday, December 1, 2010 - 2:49:26 AM

Identifiers

  • HAL Id : tel-00512748, version 1

Citation

Florian de Boissieu. Toucher artificiel à base d'un microcapteur d'effort: traitement du signal et des informations associées. Traitement du signal et de l'image [eess.SP]. Institut National Polytechnique de Grenoble - INPG, 2010. Français. ⟨tel-00512748⟩

Share

Metrics

Record views

972

Files downloads

1448