elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Imprint | Privacy Policy | Contact | Deutsch
Fontsize: [-] Text [+]

Bayesian and Neural Inference on LSTM-Based Object Recognition From Tactile and Kinesthetic Information

Pastor, Francisco and García-Gonzalez, Jorge and Gandarias, Juan M. and Medina, Daniel and Closas, Pau and García-Cerezo, Alfonso and Gómez-de-Gabriel, Jesús (2021) Bayesian and Neural Inference on LSTM-Based Object Recognition From Tactile and Kinesthetic Information. IEEE Robotics and Automation Letters, 6 (1), pp. 231-238. IEEE - Institute of Electrical and Electronics Engineers. doi: 10.1109/LRA.2020.3038377. ISSN 2377-3766.

[img] PDF - Postprint version (accepted manuscript)
1MB

Official URL: https://ieeexplore.ieee.org/document/9261100

Abstract

Recent advances in the field of intelligent robotic manipulation pursue providing robotic hands with touch sensitivity. Haptic perception encompasses the sensing modalities encountered in the sense of touch (e.g., tactile and kinesthetic sensations). This letter focuses on multimodal object recognition and proposes analytical and data-driven methodologies to fuse tactileand kinesthetic-based classification results. The procedure is as follows: a three-finger actuated gripper with an integrated high resolution tactile sensor performs squeeze-and-release Exploratory Procedures (EPs). The tactile images and kinesthetic information acquired using angular sensors on the finger joints constitute the time-series datasets of interest. Each temporal dataset is fed to a Long Short-term Memory (LSTM) Neural Network, which is trained to classify in-hand objects. The LSTMs provide an estimation of the posterior probability of each object given the corresponding measurements, which after fusion allows to estimate the object through Bayesian and Neural inference approaches. An experiment with 36-classes is carried out to evaluate and compare the performance of the fused, tactile, and kinesthetic perception systems. The results show that the Bayesian-based classifiers improves capabilities for object recognition and outperforms the Neural-based approach.

Item URL in elib:https://elib.dlr.de/139101/
Document Type:Article
Title:Bayesian and Neural Inference on LSTM-Based Object Recognition From Tactile and Kinesthetic Information
Authors:
AuthorsInstitution or Email of AuthorsAuthor's ORCID iD
Pastor, Franciscofpastor (at) uma.esUNSPECIFIED
García-Gonzalez, Jorgejorgegarcia (at) lcc.uma.esUNSPECIFIED
Gandarias, Juan M.jmgandarias (at) uma.esUNSPECIFIED
Medina, DanielDaniel.AriasMedina (at) dlr.dehttps://orcid.org/0000-0002-1586-3269
Closas, Paupau.closas (at) northeastern.eduhttps://orcid.org/0000-0002-5960-6600
García-Cerezo, Alfonsoajgarcia (at) uma.esUNSPECIFIED
Gómez-de-Gabriel, Jesúsjesus.gomez (at) uma.esUNSPECIFIED
Date:January 2021
Journal or Publication Title:IEEE Robotics and Automation Letters
Refereed publication:Yes
Open Access:Yes
Gold Open Access:No
In SCOPUS:Yes
In ISI Web of Science:Yes
Volume:6
DOI :10.1109/LRA.2020.3038377
Page Range:pp. 231-238
Publisher:IEEE - Institute of Electrical and Electronics Engineers
ISSN:2377-3766
Status:Published
Keywords:Deep learning in grasping and manipulation; force and tactile sensing; sensor fusion;
HGF - Research field:Aeronautics, Space and Transport
HGF - Program:Space
HGF - Program Themes:Communication and Navigation
DLR - Research area:Raumfahrt
DLR - Program:R KN - Kommunikation und Navigation
DLR - Research theme (Project):R - Project Navigation 4.0 (old)
Location: Neustrelitz
Institutes and Institutions:Institute of Communication and Navigation > Nautical Systems
Deposited By: Medina, Daniel
Deposited On:04 Dec 2020 15:41
Last Modified:04 Dec 2020 15:41

Repository Staff Only: item control page

Browse
Search
Help & Contact
Information
electronic library is running on EPrints 3.3.12
Copyright © 2008-2017 German Aerospace Center (DLR). All rights reserved.