Multimodal Material Identication through Recursive Tactile Sensing

Augusto Gomez Eguiluz, Ignacio Rano, Sonya Coleman, T.Martin McGinnity

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Tactile sensing has recently been used in robotics for object identication, grasping,
and material identication. Although human tactile sensing is multimodal,
existing material recognition approaches use vibration information only. Moreover,
material identication through tactile sensing can be solved as an continuous
process, yet state of the art approaches use a batch approach where readings
are taken for at least one second. This work proposes a recursive multimodal
(vibration and thermal) tactile material identication approach. Using the frequency
response of the vibration induced by the material and a set of thermal
features, we show that it is possible to accurately identify materials in less than
half a second. We conducted an exhaustive comparison of our approach with
commonly used vibration descriptors and machine learning algorithms for material
identication such as k-Nearest Neighbour, Articial Neural Network and
Support Vector Machines. Experimental results show that our approach identi
es materials faster than existing techniques and increase the classication
accuracy when multiple sensor modalities are used.
LanguageEnglish
Pages130-139
Number of pages19
JournalRobotics and Autonomous Systems
Volume106
Early online date14 May 2018
DOIs
Publication statusE-pub ahead of print - 14 May 2018

Fingerprint

Sensing
Vibration
Grasping
Modality
Learning algorithms
Descriptors
Batch
Learning systems
Robotics
Learning Algorithm
Nearest Neighbor
Machine Learning
Neural Networks
Neural networks
Sensor
Sensors
Experimental Results

Keywords

  • Material classification
  • tactile sensing
  • supervised learning

Cite this

@article{b9b2eab3eb3f4353b084f23372427d5e,
title = "Multimodal Material Identication through Recursive Tactile Sensing",
abstract = "Tactile sensing has recently been used in robotics for object identication, grasping,and material identication. Although human tactile sensing is multimodal,existing material recognition approaches use vibration information only. Moreover,material identication through tactile sensing can be solved as an continuousprocess, yet state of the art approaches use a batch approach where readingsare taken for at least one second. This work proposes a recursive multimodal(vibration and thermal) tactile material identication approach. Using the frequencyresponse of the vibration induced by the material and a set of thermalfeatures, we show that it is possible to accurately identify materials in less thanhalf a second. We conducted an exhaustive comparison of our approach withcommonly used vibration descriptors and machine learning algorithms for materialidentication such as k-Nearest Neighbour, Articial Neural Network andSupport Vector Machines. Experimental results show that our approach identies materials faster than existing techniques and increase the classicationaccuracy when multiple sensor modalities are used.",
keywords = "Material classification, tactile sensing, supervised learning",
author = "{Gomez Eguiluz}, Augusto and Ignacio Rano and Sonya Coleman and T.Martin McGinnity",
year = "2018",
month = "5",
day = "14",
doi = "10.1016/j.robot.2018.05.003",
language = "English",
volume = "106",
pages = "130--139",
journal = "Robotics and Autonomous Systems",
issn = "0921-8890",
publisher = "Elsevier",

}

Multimodal Material Identication through Recursive Tactile Sensing. / Gomez Eguiluz, Augusto; Rano, Ignacio; Coleman, Sonya; McGinnity, T.Martin.

In: Robotics and Autonomous Systems, Vol. 106, 14.05.2018, p. 130-139.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Multimodal Material Identication through Recursive Tactile Sensing

AU - Gomez Eguiluz, Augusto

AU - Rano, Ignacio

AU - Coleman, Sonya

AU - McGinnity, T.Martin

PY - 2018/5/14

Y1 - 2018/5/14

N2 - Tactile sensing has recently been used in robotics for object identication, grasping,and material identication. Although human tactile sensing is multimodal,existing material recognition approaches use vibration information only. Moreover,material identication through tactile sensing can be solved as an continuousprocess, yet state of the art approaches use a batch approach where readingsare taken for at least one second. This work proposes a recursive multimodal(vibration and thermal) tactile material identication approach. Using the frequencyresponse of the vibration induced by the material and a set of thermalfeatures, we show that it is possible to accurately identify materials in less thanhalf a second. We conducted an exhaustive comparison of our approach withcommonly used vibration descriptors and machine learning algorithms for materialidentication such as k-Nearest Neighbour, Articial Neural Network andSupport Vector Machines. Experimental results show that our approach identies materials faster than existing techniques and increase the classicationaccuracy when multiple sensor modalities are used.

AB - Tactile sensing has recently been used in robotics for object identication, grasping,and material identication. Although human tactile sensing is multimodal,existing material recognition approaches use vibration information only. Moreover,material identication through tactile sensing can be solved as an continuousprocess, yet state of the art approaches use a batch approach where readingsare taken for at least one second. This work proposes a recursive multimodal(vibration and thermal) tactile material identication approach. Using the frequencyresponse of the vibration induced by the material and a set of thermalfeatures, we show that it is possible to accurately identify materials in less thanhalf a second. We conducted an exhaustive comparison of our approach withcommonly used vibration descriptors and machine learning algorithms for materialidentication such as k-Nearest Neighbour, Articial Neural Network andSupport Vector Machines. Experimental results show that our approach identies materials faster than existing techniques and increase the classicationaccuracy when multiple sensor modalities are used.

KW - Material classification

KW - tactile sensing

KW - supervised learning

U2 - 10.1016/j.robot.2018.05.003

DO - 10.1016/j.robot.2018.05.003

M3 - Article

VL - 106

SP - 130

EP - 139

JO - Robotics and Autonomous Systems

T2 - Robotics and Autonomous Systems

JF - Robotics and Autonomous Systems

SN - 0921-8890

ER -