Control of a Soft Actuator using a Long Short-Term Memory Neural Network

Victor Yanev; Maria Elena Giannaccini; Sumeet S. Aphale

doi:10.1109/MMAR55195.2022.9874322

Control of a Soft Actuator using a Long Short-Term Memory Neural Network

Victor Yanev, Maria Elena Giannaccini, Sumeet S. Aphale

University of Aberdeen

Research output: Chapter in Book/Report/Conference proceeding › Published conference contribution

1 Citation (Scopus)

Abstract

Soft robots offer new opportunities because of their compliant physical structure and their wide range of applications. Currently the development of such robots is hampered by their low controllability. One of the main constituents of soft robots are soft actuators. The aim of this project is to improve the control of a non-linear system, the soft actuator, and its interaction with the environment, by training a long short-term memory (LSTM) neural network to accurately predict the actuator's position in space, its curvature, and the force applied by its end-effector on an external object. The increased performance of the trained network resulted in an error as low as 0.01±0.005N in estimating the force applied by the end effector on the external object. The results show significantly superior performance (on the order of 10 times) in the positional and curvature predictions of the LSTM network when using one marker per air-chamber.

Original language	English
Title of host publication	2022 26th International Conference on Methods and Models in Automation and Robotics
Subtitle of host publication	MMAR 2022 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	187-192
Number of pages	6
ISBN (Electronic)	9781665468572
DOIs	https://doi.org/10.1109/MMAR55195.2022.9874322
Publication status	Published - 8 Sept 2022
Event	26th International Conference on Methods and Models in Automation and Robotics, MMAR 2022 - Virtual, Miedzyzdroje, Poland Duration: 22 Aug 2022 → 25 Aug 2022

Conference

Conference	26th International Conference on Methods and Models in Automation and Robotics, MMAR 2022
Country/Territory	Poland
City	Virtual, Miedzyzdroje
Period	22/08/22 → 25/08/22

Bibliographical note

This work was supported by the Carnegie Trust Vacation Scholarship funding, awarded to Victor Yanev. All of the code written for this paper is available and can be enquired by emailing one of the authors.

Keywords

control
LSTM
neural networks
soft pneumatic actuator
soft robotics

Access to Document

10.1109/MMAR55195.2022.9874322Licence: Unspecified

Cite this

Control of a Soft Actuator using a Long Short-Term Memory Neural Network. / Yanev, Victor; Giannaccini, Maria Elena ; Aphale, Sumeet S.
2022 26th International Conference on Methods and Models in Automation and Robotics: MMAR 2022 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2022. p. 187-192.

Research output: Chapter in Book/Report/Conference proceeding › Published conference contribution

Yanev, V, Giannaccini, ME & Aphale, SS 2022, Control of a Soft Actuator using a Long Short-Term Memory Neural Network. in 2022 26th International Conference on Methods and Models in Automation and Robotics: MMAR 2022 - Proceedings. Institute of Electrical and Electronics Engineers Inc., pp. 187-192, 26th International Conference on Methods and Models in Automation and Robotics, MMAR 2022, Virtual, Miedzyzdroje, Poland, 22/08/22. https://doi.org/10.1109/MMAR55195.2022.9874322

@inproceedings{4349715e8c8647869ee1b36d844085d3,

title = "Control of a Soft Actuator using a Long Short-Term Memory Neural Network",

abstract = "Soft robots offer new opportunities because of their compliant physical structure and their wide range of applications. Currently the development of such robots is hampered by their low controllability. One of the main constituents of soft robots are soft actuators. The aim of this project is to improve the control of a non-linear system, the soft actuator, and its interaction with the environment, by training a long short-term memory (LSTM) neural network to accurately predict the actuator's position in space, its curvature, and the force applied by its end-effector on an external object. The increased performance of the trained network resulted in an error as low as 0.01±0.005N in estimating the force applied by the end effector on the external object. The results show significantly superior performance (on the order of 10 times) in the positional and curvature predictions of the LSTM network when using one marker per air-chamber.",

keywords = "control, LSTM, neural networks, soft pneumatic actuator, soft robotics",

author = "Victor Yanev and Giannaccini, {Maria Elena} and Aphale, {Sumeet S.}",

note = "This work was supported by the Carnegie Trust Vacation Scholarship funding, awarded to Victor Yanev. All of the code written for this paper is available and can be enquired by emailing one of the authors. ; 26th International Conference on Methods and Models in Automation and Robotics, MMAR 2022 ; Conference date: 22-08-2022 Through 25-08-2022",

year = "2022",

month = sep,

day = "8",

doi = "10.1109/MMAR55195.2022.9874322",

language = "English",

pages = "187--192",

booktitle = "2022 26th International Conference on Methods and Models in Automation and Robotics",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

address = "United States",

}

TY - GEN

T1 - Control of a Soft Actuator using a Long Short-Term Memory Neural Network

AU - Yanev, Victor

AU - Giannaccini, Maria Elena

AU - Aphale, Sumeet S.

N1 - This work was supported by the Carnegie Trust Vacation Scholarship funding, awarded to Victor Yanev. All of the code written for this paper is available and can be enquired by emailing one of the authors.

PY - 2022/9/8

Y1 - 2022/9/8

N2 - Soft robots offer new opportunities because of their compliant physical structure and their wide range of applications. Currently the development of such robots is hampered by their low controllability. One of the main constituents of soft robots are soft actuators. The aim of this project is to improve the control of a non-linear system, the soft actuator, and its interaction with the environment, by training a long short-term memory (LSTM) neural network to accurately predict the actuator's position in space, its curvature, and the force applied by its end-effector on an external object. The increased performance of the trained network resulted in an error as low as 0.01±0.005N in estimating the force applied by the end effector on the external object. The results show significantly superior performance (on the order of 10 times) in the positional and curvature predictions of the LSTM network when using one marker per air-chamber.

AB - Soft robots offer new opportunities because of their compliant physical structure and their wide range of applications. Currently the development of such robots is hampered by their low controllability. One of the main constituents of soft robots are soft actuators. The aim of this project is to improve the control of a non-linear system, the soft actuator, and its interaction with the environment, by training a long short-term memory (LSTM) neural network to accurately predict the actuator's position in space, its curvature, and the force applied by its end-effector on an external object. The increased performance of the trained network resulted in an error as low as 0.01±0.005N in estimating the force applied by the end effector on the external object. The results show significantly superior performance (on the order of 10 times) in the positional and curvature predictions of the LSTM network when using one marker per air-chamber.

KW - control

KW - LSTM

KW - neural networks

KW - soft pneumatic actuator

KW - soft robotics

UR - http://www.scopus.com/inward/record.url?scp=85139036712&partnerID=8YFLogxK

U2 - 10.1109/MMAR55195.2022.9874322

DO - 10.1109/MMAR55195.2022.9874322

M3 - Published conference contribution

AN - SCOPUS:85139036712

SP - 187

EP - 192

BT - 2022 26th International Conference on Methods and Models in Automation and Robotics

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 26th International Conference on Methods and Models in Automation and Robotics, MMAR 2022

Y2 - 22 August 2022 through 25 August 2022

ER -

Control of a Soft Actuator using a Long Short-Term Memory Neural Network

Abstract

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this