Comparison of Deep Neural Network Models and Effectiveness of EMG Signal Feature Value for Estimating Dorsiflexion

Robotic ankle–foot orthoses (AFO) are often used for gait rehabilitation. Our research focuses on the design and development of a robotic AFO with minimum number of sensor inputs. However, this leads to degradation of gait estimation accuracy. To prevent degradation of accuracy, we compared a few ne...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Muhammad Akmal Bin Mohammed Zaffir, Praveen Nuwantha, Daiki Arase, Keiko Sakurai, Hiroki Tamura
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	neural network ANFIS MLP LSTM deep learning AFOs Electronics TK7800-8360
Acceso en línea:	https://doaj.org/article/820c12bef80a4ef5a1b74f36b38467e0
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:820c12bef80a4ef5a1b74f36b38467e0
record_format	dspace
spelling	oai:doaj.org-article:820c12bef80a4ef5a1b74f36b38467e02021-11-25T17:24:28ZComparison of Deep Neural Network Models and Effectiveness of EMG Signal Feature Value for Estimating Dorsiflexion10.3390/electronics102227672079-9292https://doaj.org/article/820c12bef80a4ef5a1b74f36b38467e02021-11-01T00:00:00Zhttps://www.mdpi.com/2079-9292/10/22/2767https://doaj.org/toc/2079-9292Robotic ankle–foot orthoses (AFO) are often used for gait rehabilitation. Our research focuses on the design and development of a robotic AFO with minimum number of sensor inputs. However, this leads to degradation of gait estimation accuracy. To prevent degradation of accuracy, we compared a few neural network models in order to determine the best network when only two input channels are being used. Further, the EMG signal feature value of average rate of change was used as input. LSTM showed the highest accuracy. However, MLP with a small number of hidden layers showed results similar to LSTM. Moreover, the accuracy for all models, with the exception of LSTM for one subject (SD), increased with the addition of feature value (average rate of change) as input. In conclusion, time-series networks work best with a small number of sensor inputs. However, depending on the optimizer being used, even a simple network can outrun a deep learning network. Furthermore, our results show that applying EMG signal feature value as an input tends to increase the estimation accuracy of the network.Muhammad Akmal Bin Mohammed ZaffirPraveen NuwanthaDaiki AraseKeiko SakuraiHiroki TamuraMDPI AGarticleneural networkANFISMLPLSTMdeep learningAFOsElectronicsTK7800-8360ENElectronics, Vol 10, Iss 2767, p 2767 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	neural network ANFIS MLP LSTM deep learning AFOs Electronics TK7800-8360
spellingShingle	neural network ANFIS MLP LSTM deep learning AFOs Electronics TK7800-8360 Muhammad Akmal Bin Mohammed Zaffir Praveen Nuwantha Daiki Arase Keiko Sakurai Hiroki Tamura Comparison of Deep Neural Network Models and Effectiveness of EMG Signal Feature Value for Estimating Dorsiflexion
description	Robotic ankle–foot orthoses (AFO) are often used for gait rehabilitation. Our research focuses on the design and development of a robotic AFO with minimum number of sensor inputs. However, this leads to degradation of gait estimation accuracy. To prevent degradation of accuracy, we compared a few neural network models in order to determine the best network when only two input channels are being used. Further, the EMG signal feature value of average rate of change was used as input. LSTM showed the highest accuracy. However, MLP with a small number of hidden layers showed results similar to LSTM. Moreover, the accuracy for all models, with the exception of LSTM for one subject (SD), increased with the addition of feature value (average rate of change) as input. In conclusion, time-series networks work best with a small number of sensor inputs. However, depending on the optimizer being used, even a simple network can outrun a deep learning network. Furthermore, our results show that applying EMG signal feature value as an input tends to increase the estimation accuracy of the network.
format	article
author	Muhammad Akmal Bin Mohammed Zaffir Praveen Nuwantha Daiki Arase Keiko Sakurai Hiroki Tamura
author_facet	Muhammad Akmal Bin Mohammed Zaffir Praveen Nuwantha Daiki Arase Keiko Sakurai Hiroki Tamura
author_sort	Muhammad Akmal Bin Mohammed Zaffir
title	Comparison of Deep Neural Network Models and Effectiveness of EMG Signal Feature Value for Estimating Dorsiflexion
title_short	Comparison of Deep Neural Network Models and Effectiveness of EMG Signal Feature Value for Estimating Dorsiflexion
title_full	Comparison of Deep Neural Network Models and Effectiveness of EMG Signal Feature Value for Estimating Dorsiflexion
title_fullStr	Comparison of Deep Neural Network Models and Effectiveness of EMG Signal Feature Value for Estimating Dorsiflexion
title_full_unstemmed	Comparison of Deep Neural Network Models and Effectiveness of EMG Signal Feature Value for Estimating Dorsiflexion
title_sort	comparison of deep neural network models and effectiveness of emg signal feature value for estimating dorsiflexion
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/820c12bef80a4ef5a1b74f36b38467e0
work_keys_str_mv	AT muhammadakmalbinmohammedzaffir comparisonofdeepneuralnetworkmodelsandeffectivenessofemgsignalfeaturevalueforestimatingdorsiflexion AT praveennuwantha comparisonofdeepneuralnetworkmodelsandeffectivenessofemgsignalfeaturevalueforestimatingdorsiflexion AT daikiarase comparisonofdeepneuralnetworkmodelsandeffectivenessofemgsignalfeaturevalueforestimatingdorsiflexion AT keikosakurai comparisonofdeepneuralnetworkmodelsandeffectivenessofemgsignalfeaturevalueforestimatingdorsiflexion AT hirokitamura comparisonofdeepneuralnetworkmodelsandeffectivenessofemgsignalfeaturevalueforestimatingdorsiflexion
_version_	1718412417940062208

Comparison of Deep Neural Network Models and Effectiveness of EMG Signal Feature Value for Estimating Dorsiflexion

Ejemplares similares