Development of EMG measurement system aimed to control assisted device

Akihiko Hanafusa, Tomohiko Shimizu

研究成果: Article

抄録

The objective of the study is to develop an EMG measurement system aimed to control an active assisted device that is used by a subject whose upper limb is paralyzed as Erb’s palsy. The EMG measurement circuit with eight channel differential amplifiers, rectifier and filter and the software that discriminate the motion were developed. For discriminant method, nearest neighbor algorithm using Euclidean (EUC) and Mahalanobis (MH) distances, and support vector machine (SVM) were used. The EMG by six motions with radial and ulnar flexion, extension-flexion of wrist were and open-close movement of fingers were measured and evaluated by 4, 6 and 8 channels. The discriminant ratio was lowest by four channels; however the result of six channels was almost same as that of eight when the appropriate channels were selected. 100% discriminant ratio can be obtained by choosing the largest number of distinguished motion using SVM and EUC methods.

元の言語English
ページ(範囲)O-32-O-33
ジャーナルTransactions of Japanese Society for Medical and Biological Engineering
52
DOI
出版物ステータスPublished - 2014 8 17

Fingerprint

Support vector machines
Differential amplifiers
Networks (circuits)

ASJC Scopus subject areas

  • Biomedical Engineering

これを引用

@article{a7b3a611c7d348918c763c5d8ebcea02,
title = "Development of EMG measurement system aimed to control assisted device",
abstract = "The objective of the study is to develop an EMG measurement system aimed to control an active assisted device that is used by a subject whose upper limb is paralyzed as Erb’s palsy. The EMG measurement circuit with eight channel differential amplifiers, rectifier and filter and the software that discriminate the motion were developed. For discriminant method, nearest neighbor algorithm using Euclidean (EUC) and Mahalanobis (MH) distances, and support vector machine (SVM) were used. The EMG by six motions with radial and ulnar flexion, extension-flexion of wrist were and open-close movement of fingers were measured and evaluated by 4, 6 and 8 channels. The discriminant ratio was lowest by four channels; however the result of six channels was almost same as that of eight when the appropriate channels were selected. 100{\%} discriminant ratio can be obtained by choosing the largest number of distinguished motion using SVM and EUC methods.",
keywords = "Assisted Device, Discriminant Method, EMG",
author = "Akihiko Hanafusa and Tomohiko Shimizu",
year = "2014",
month = "8",
day = "17",
doi = "10.11239/jsmbe.52.O-32",
language = "English",
volume = "52",
pages = "O--32--O--33",
journal = "BME = Bio medical engineering / henshu, Nihon ME Gakkai",
issn = "1347-443X",
publisher = "Nihon M E Gakkai",

}

TY - JOUR

T1 - Development of EMG measurement system aimed to control assisted device

AU - Hanafusa, Akihiko

AU - Shimizu, Tomohiko

PY - 2014/8/17

Y1 - 2014/8/17

N2 - The objective of the study is to develop an EMG measurement system aimed to control an active assisted device that is used by a subject whose upper limb is paralyzed as Erb’s palsy. The EMG measurement circuit with eight channel differential amplifiers, rectifier and filter and the software that discriminate the motion were developed. For discriminant method, nearest neighbor algorithm using Euclidean (EUC) and Mahalanobis (MH) distances, and support vector machine (SVM) were used. The EMG by six motions with radial and ulnar flexion, extension-flexion of wrist were and open-close movement of fingers were measured and evaluated by 4, 6 and 8 channels. The discriminant ratio was lowest by four channels; however the result of six channels was almost same as that of eight when the appropriate channels were selected. 100% discriminant ratio can be obtained by choosing the largest number of distinguished motion using SVM and EUC methods.

AB - The objective of the study is to develop an EMG measurement system aimed to control an active assisted device that is used by a subject whose upper limb is paralyzed as Erb’s palsy. The EMG measurement circuit with eight channel differential amplifiers, rectifier and filter and the software that discriminate the motion were developed. For discriminant method, nearest neighbor algorithm using Euclidean (EUC) and Mahalanobis (MH) distances, and support vector machine (SVM) were used. The EMG by six motions with radial and ulnar flexion, extension-flexion of wrist were and open-close movement of fingers were measured and evaluated by 4, 6 and 8 channels. The discriminant ratio was lowest by four channels; however the result of six channels was almost same as that of eight when the appropriate channels were selected. 100% discriminant ratio can be obtained by choosing the largest number of distinguished motion using SVM and EUC methods.

KW - Assisted Device

KW - Discriminant Method

KW - EMG

UR - http://www.scopus.com/inward/record.url?scp=84939422667&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84939422667&partnerID=8YFLogxK

U2 - 10.11239/jsmbe.52.O-32

DO - 10.11239/jsmbe.52.O-32

M3 - Article

AN - SCOPUS:84939422667

VL - 52

SP - O-32-O-33

JO - BME = Bio medical engineering / henshu, Nihon ME Gakkai

JF - BME = Bio medical engineering / henshu, Nihon ME Gakkai

SN - 1347-443X

ER -