Optimization of neural network hyperparameters for modeling of soft pneumatic actuators

Research output: Chapter in book/report/conference proceedingContribution to book/anthologyResearchpeer review

Authors

View graph of relations

Details

Original languageEnglish
Title of host publicationMechanisms and Machine Science
PublisherSpringer Netherlands
Pages199-206
Number of pages8
Volume65
Publication statusPublished - 27 Sept 2018

Publication series

NameMechanisms and Machine Science
Volume65
ISSN (Print)2211-0984
ISSN (electronic)2211-0992

Abstract

Especially the field of medical robotics is faced with the challenge that contact between a robotic structure is not only tolerated but at the very core of the application, for example in minimally invasive surgery. Therefore, high demands regarding safety need to be met when operating a robotic structure in such a delicate environment. Soft material robotic systems offer the potential to surpass their rigid counterparts in this area due to their material inherent compliance and adaptability. Nonetheless, soft structures can also exhibit challenging characteristics like highly nonlinear deformation. This effect is often neglected when setting up deformation models. Recent approaches tackle this challenge by applying machine learning methods to learn the nonlinear behaviour. In previous research, we gained promising results from learning the kinematics of a soft pneumatic actuator (SPA) via a feedforward artificial neural network (ANN). To overcome a trial and error approach when designing an ANN, in this article we introduce Bayesian optimization to find suitable hyperparameters. This way an ANN architecture is found that exceeds the accuracy of our previous studies by more than a factor of 20 [9].

Keywords

    Artificial neural networks, Hyperparameter optimization, Kinematic modeling, Soft material robotic systems, Soft pneumatic actuators

ASJC Scopus subject areas

Cite this

Optimization of neural network hyperparameters for modeling of soft pneumatic actuators. / Wiese, Mats; Runge-Borchert, Gundula; Raatz, Annika.
Mechanisms and Machine Science. Vol. 65 Springer Netherlands, 2018. p. 199-206 (Mechanisms and Machine Science; Vol. 65).

Research output: Chapter in book/report/conference proceedingContribution to book/anthologyResearchpeer review

Wiese, M, Runge-Borchert, G & Raatz, A 2018, Optimization of neural network hyperparameters for modeling of soft pneumatic actuators. in Mechanisms and Machine Science. vol. 65, Mechanisms and Machine Science, vol. 65, Springer Netherlands, pp. 199-206. https://doi.org/10.1007/978-3-030-00329-6_23
Wiese, M., Runge-Borchert, G., & Raatz, A. (2018). Optimization of neural network hyperparameters for modeling of soft pneumatic actuators. In Mechanisms and Machine Science (Vol. 65, pp. 199-206). (Mechanisms and Machine Science; Vol. 65). Springer Netherlands. https://doi.org/10.1007/978-3-030-00329-6_23
Wiese M, Runge-Borchert G, Raatz A. Optimization of neural network hyperparameters for modeling of soft pneumatic actuators. In Mechanisms and Machine Science. Vol. 65. Springer Netherlands. 2018. p. 199-206. (Mechanisms and Machine Science). doi: 10.1007/978-3-030-00329-6_23
Wiese, Mats ; Runge-Borchert, Gundula ; Raatz, Annika. / Optimization of neural network hyperparameters for modeling of soft pneumatic actuators. Mechanisms and Machine Science. Vol. 65 Springer Netherlands, 2018. pp. 199-206 (Mechanisms and Machine Science).
Download
@inbook{75b2614bad9944bfa2253ddcab50db9e,
title = "Optimization of neural network hyperparameters for modeling of soft pneumatic actuators",
abstract = "Especially the field of medical robotics is faced with the challenge that contact between a robotic structure is not only tolerated but at the very core of the application, for example in minimally invasive surgery. Therefore, high demands regarding safety need to be met when operating a robotic structure in such a delicate environment. Soft material robotic systems offer the potential to surpass their rigid counterparts in this area due to their material inherent compliance and adaptability. Nonetheless, soft structures can also exhibit challenging characteristics like highly nonlinear deformation. This effect is often neglected when setting up deformation models. Recent approaches tackle this challenge by applying machine learning methods to learn the nonlinear behaviour. In previous research, we gained promising results from learning the kinematics of a soft pneumatic actuator (SPA) via a feedforward artificial neural network (ANN). To overcome a trial and error approach when designing an ANN, in this article we introduce Bayesian optimization to find suitable hyperparameters. This way an ANN architecture is found that exceeds the accuracy of our previous studies by more than a factor of 20 [9].",
keywords = "Artificial neural networks, Hyperparameter optimization, Kinematic modeling, Soft material robotic systems, Soft pneumatic actuators",
author = "Mats Wiese and Gundula Runge-Borchert and Annika Raatz",
year = "2018",
month = sep,
day = "27",
doi = "10.1007/978-3-030-00329-6_23",
language = "English",
volume = "65",
series = "Mechanisms and Machine Science",
publisher = "Springer Netherlands",
pages = "199--206",
booktitle = "Mechanisms and Machine Science",
address = "Netherlands",

}

Download

TY - CHAP

T1 - Optimization of neural network hyperparameters for modeling of soft pneumatic actuators

AU - Wiese, Mats

AU - Runge-Borchert, Gundula

AU - Raatz, Annika

PY - 2018/9/27

Y1 - 2018/9/27

N2 - Especially the field of medical robotics is faced with the challenge that contact between a robotic structure is not only tolerated but at the very core of the application, for example in minimally invasive surgery. Therefore, high demands regarding safety need to be met when operating a robotic structure in such a delicate environment. Soft material robotic systems offer the potential to surpass their rigid counterparts in this area due to their material inherent compliance and adaptability. Nonetheless, soft structures can also exhibit challenging characteristics like highly nonlinear deformation. This effect is often neglected when setting up deformation models. Recent approaches tackle this challenge by applying machine learning methods to learn the nonlinear behaviour. In previous research, we gained promising results from learning the kinematics of a soft pneumatic actuator (SPA) via a feedforward artificial neural network (ANN). To overcome a trial and error approach when designing an ANN, in this article we introduce Bayesian optimization to find suitable hyperparameters. This way an ANN architecture is found that exceeds the accuracy of our previous studies by more than a factor of 20 [9].

AB - Especially the field of medical robotics is faced with the challenge that contact between a robotic structure is not only tolerated but at the very core of the application, for example in minimally invasive surgery. Therefore, high demands regarding safety need to be met when operating a robotic structure in such a delicate environment. Soft material robotic systems offer the potential to surpass their rigid counterparts in this area due to their material inherent compliance and adaptability. Nonetheless, soft structures can also exhibit challenging characteristics like highly nonlinear deformation. This effect is often neglected when setting up deformation models. Recent approaches tackle this challenge by applying machine learning methods to learn the nonlinear behaviour. In previous research, we gained promising results from learning the kinematics of a soft pneumatic actuator (SPA) via a feedforward artificial neural network (ANN). To overcome a trial and error approach when designing an ANN, in this article we introduce Bayesian optimization to find suitable hyperparameters. This way an ANN architecture is found that exceeds the accuracy of our previous studies by more than a factor of 20 [9].

KW - Artificial neural networks

KW - Hyperparameter optimization

KW - Kinematic modeling

KW - Soft material robotic systems

KW - Soft pneumatic actuators

UR - http://www.scopus.com/inward/record.url?scp=85054128819&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-00329-6_23

DO - 10.1007/978-3-030-00329-6_23

M3 - Contribution to book/anthology

AN - SCOPUS:85054128819

VL - 65

T3 - Mechanisms and Machine Science

SP - 199

EP - 206

BT - Mechanisms and Machine Science

PB - Springer Netherlands

ER -

By the same author(s)