Article(id=1245407864299303881, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1156262727438951343, articleNumber=null, orderNo=null, doi=10.12404/j.issn.1671-1815.2403062, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=research-article, receivedDate=1713974400000, receivedDateStr=2024-04-25, revisedDate=1717516800000, revisedDateStr=2024-06-05, acceptedDate=null, acceptedDateStr=null, onlineDate=1774857973408, onlineDateStr=2026-03-30, pubDate=1741363200000, pubDateStr=2025-03-08, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1774857973408, onlineIssueDateStr=2026-03-30, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1774857973408, creator=13701087609, updateTime=1774857973408, updator=13701087609, issue=Issue{id=1156262727438951343, tenantId=1146029695717560320, journalId=1146123166801305609, year='2025', volume='25', issue='7', pageStart='2193', pageEnd='3077', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=0, createTime=1753604116544, creator=13701087609, updateTime=1753771263994, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1156963794699248405, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1156262727438951343, language=EN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1156963794699248406, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1156262727438951343, language=CN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=2841, endPage=2848, ext={EN=ArticleExt(id=1245407864865533957, articleId=1245407864299303881, tenantId=1146029695717560320, journalId=1146123166801305609, language=EN, title=Lower Limb Motion Recognition Based on Surface Electromyography and a CNN-LSTM Fusion Model, columnId=1156262729162810294, journalTitle=Science Technology and Engineering, columnName=Papers·Automation and Computational Technology, runingTitle=null, highlight=null, articleAbstract=

To enhance the classification accuracy of lower limb movements, this paper was introduced a hybrid recognition model based on surface electromyography (sEMG) that combines convolutional neural networks (CNN) with long short-term memory networks (LSTM). Initially, sEMG data were collected from 20 subjects performing four types of gait movements: ascending stairs, descending stairs, walking, and squatting. Subsequently, the collected sEMG data underwent preprocessing, and both time domain and frequency domain features were extracted to serve as inputs for the machine learning recognition model. The CNN-LSTM model was then constructed for lower limb action recognition and compared against the performances of CNN, LSTM, and SVM (support vector machine,)models. The results demonstrate that the CNN-LSTM model outperforms the CNN, LSTM, and SVM models by 2.16%, 8.34%, and 11.16% in accuracy, respectively, thereby proving its superior classification performance. This model provides an effective solution for enhancing lower limb motor functions, offering significant benefits for rehabilitation medical equipment and power assist devices.

, correspAuthors=Qing TAO, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, authorCompany=null, fund=null, authors=null, authorsList=Zhi-wei ZHOU, Qing TAO, Na SU, Jing-xuan LIU, Bo-wen LI, Hao PEI), CN=ArticleExt(id=1245407871136018757, articleId=1245407864299303881, tenantId=1146029695717560320, journalId=1146123166801305609, language=CN, title=基于表面肌电信号的CNN-LSTM模型下肢动作识别, columnId=1156262729783567290, journalTitle=科学技术与工程, columnName=论文·自动化技术、计算机技术, runingTitle=null, highlight=null, articleAbstract=

为了提高对下肢运动的分类准确度,提出了一种基于表面肌电信号(surface electromyography, sEMG)的卷积神经网络与长短期记忆网络融合识别模型(convolutional neural network and long short-term memory network, CNN-LSTM)。首先,采集了20名受试者进行上楼、下楼、行走和蹲起4种步态动作的sEMG;接着,对采集到的sEMG数据进行预处理,并提取了两种时域和频域特征,用作机器学习识别模型的特征输入;最后,基于预处理后肌电信号数据,构建了CNN-LSTM的下肢动作识别模型,并与CNN、LSTM和支持向量机(support vector machine, SVM)模型的性能进行对比。结果显示,CNN-LSTM模型在下肢动作识别准确率上分别比CNN、LSTM和SVM模型高出2.16%、8.34%、和11.16%,证明了其优越的分类性能。研究结论为康复医疗器械与助力器械提供了一个有效的下肢运动功能改善方案。

, correspAuthors=陶庆, authorNote=null, correspAuthorsNote=
* 陶庆(1978—),男,汉族,江苏常熟人,博士,教授。研究方向:先进设计与制造、虚拟现实技术。E-mail:
, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=9Af0fnKAm4j4ggoTPKla2A==, magXml=zfzDlHKk7mDp0xpMuemkDA==, pdfUrl=null, pdf=8FoAi49UtDlzpzNQNHR0YA==, pdfFileSize=6367376, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=AMoDOmZkfPbW8k41X2R+yA==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=MCc2fr1xi0M5GXQQzRrVnQ==, mapNumber=null, authorCompany=null, fund=null, authors=

周智伟(1999—),男,汉族,新疆奎屯人,硕士研究生。研究方向:下肢动作识别、深度学习。E-mail:

, authorsList=周智伟, 陶庆, 苏娜, 刘景轩, 李博文, 裴浩)}, authors=[Author(id=1245407871689666954, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=912918131@qq.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1245407871832273311, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407871689666954, language=EN, stringName=Zhi-wei ZHOU, firstName=Zhi-wei, middleName=null, lastName=ZHOU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1245407871949713844, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407871689666954, language=CN, stringName=周智伟, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017, bio={"content":"

周智伟(1999—),男,汉族,新疆奎屯人,硕士研究生。研究方向:下肢动作识别、深度学习。E-mail:

"}, bioImg=null, bioContent=

周智伟(1999—),男,汉族,新疆奎屯人,硕士研究生。研究方向:下肢动作识别、深度学习。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1245407871408648552, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, xref=1, ext=[AuthorCompanyExt(id=1245407871417037162, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China), AuthorCompanyExt(id=1245407871421231467, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017)])]), Author(id=1245407872050377153, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=xjutao@qq.com, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1245407872192983508, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407872050377153, language=EN, stringName=Qing TAO, firstName=Qing, middleName=null, lastName=TAO, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, *, address=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1245407872285258212, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407872050377153, language=CN, stringName=陶庆, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, *, address=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1245407871408648552, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, xref=1, ext=[AuthorCompanyExt(id=1245407871417037162, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China), AuthorCompanyExt(id=1245407871421231467, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017)])]), Author(id=1245407872415281655, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1245407872520139269, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407872415281655, language=EN, stringName=Na SU, firstName=Na, middleName=null, lastName=SU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2 Medical Laboratory Center, The First Affiliated Hospital of Xinjiang Medical University, Urumqi 830054, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1245407872662745617, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407872415281655, language=CN, stringName=苏娜, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2 新疆医科大学第一附属医院, 医学检验中心, 乌鲁木齐 830054, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1245407871568032125, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, xref=2, ext=[AuthorCompanyExt(id=1245407871580615040, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871568032125, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 Medical Laboratory Center, The First Affiliated Hospital of Xinjiang Medical University, Urumqi 830054, China), AuthorCompanyExt(id=1245407871597392257, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871568032125, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 新疆医科大学第一附属医院, 医学检验中心, 乌鲁木齐 830054)])]), Author(id=1245407872813740578, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1245407872943764019, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407872813740578, language=EN, stringName=Jing-xuan LIU, firstName=Jing-xuan, middleName=null, lastName=LIU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1245407873094758981, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407872813740578, language=CN, stringName=刘景轩, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1245407871408648552, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, xref=1, ext=[AuthorCompanyExt(id=1245407871417037162, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China), AuthorCompanyExt(id=1245407871421231467, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017)])]), Author(id=1245407873241559640, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1245407873371583077, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407873241559640, language=EN, stringName=Bo-wen LI, firstName=Bo-wen, middleName=null, lastName=LI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1245407873509995124, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407873241559640, language=CN, stringName=李博文, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1245407871408648552, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, xref=1, ext=[AuthorCompanyExt(id=1245407871417037162, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China), AuthorCompanyExt(id=1245407871421231467, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017)])]), Author(id=1245407873644212866, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, orderNo=5, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1245407873757459086, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407873644212866, language=EN, stringName=Hao PEI, firstName=Hao, middleName=null, lastName=PEI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1245407873916842652, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, authorId=1245407873644212866, language=CN, stringName=裴浩, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1245407871408648552, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, xref=1, ext=[AuthorCompanyExt(id=1245407871417037162, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China), AuthorCompanyExt(id=1245407871421231467, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017)])])], keywords=[Keyword(id=1245407874055254697, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, orderNo=1, keyword=surface electromyographic signals), Keyword(id=1245407874252386997, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, orderNo=2, keyword=lower limb motion recognition), Keyword(id=1245407874382410436, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, orderNo=3, keyword=CNN-LSTM), Keyword(id=1245407874491462351, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, orderNo=4, keyword=convolutional neural networks), Keyword(id=1245407874625680092, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, orderNo=5, keyword=long short-term memory networks), Keyword(id=1245407874785063658, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, orderNo=1, keyword=表面肌电信号), Keyword(id=1245407874940252917, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, orderNo=2, keyword=下肢动作识别), Keyword(id=1245407875070276358, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, orderNo=3, keyword=CNN-LSTM), Keyword(id=1245407875217077011, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, orderNo=4, keyword=卷积神经网络), Keyword(id=1245407875321934627, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, orderNo=5, keyword=长短时记忆网络)], refs=[Reference(id=1245407879243608167, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2019, volume=9, issue=9, pageStart=095042, pageEnd=null, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=Yang C, Xi X G, Chen S J, journalName=AIP Advances, refType=null, unstructuredReference=Yang C, Xi X G, Chen S J, et al. SEMG-based multifeatures and predictive model for knee-joint-angle estimation[J]. AIP Advances, 2019, 9(9): 095042., articleTitle=SEMG-based multifeatures and predictive model for knee-joint-angle estimation, refAbstract=null), Reference(id=1245407879340077168, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2018, volume=65, issue=null, pageStart=1069, pageEnd=1076, url=null, language=null, rfNumber=[2], rfOrder=1, authorNames=Totty M S, Wade E, journalName=IEEE Transactions on Biomedical Engineering, refType=null, unstructuredReference=Totty M S, Wade E. Muscle activation and inertial motion data for noninvasive classification of activities of daily living[J]. IEEE Transactions on Biomedical Engineering, 2018, 65: 1069-1076., articleTitle=Muscle activation and inertial motion data for noninvasive classification of activities of daily living, refAbstract=null), Reference(id=1245407879457517687, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=35, pageStart=15133, pageEnd=15141, url=null, language=null, rfNumber=[3], rfOrder=2, authorNames=崔冰艳, 邓嘉, 张祥, journalName=科学技术与工程, refType=null, unstructuredReference=崔冰艳, 邓嘉, 张祥. 基于表面肌电信号的手势识别方法[J]. 科学技术与工程, 2023, 23(35): 15133-15141., articleTitle=基于表面肌电信号的手势识别方法, refAbstract=null), Reference(id=1245407879558180991, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=35, pageStart=15133, pageEnd=15141, url=null, language=null, rfNumber=[3], rfOrder=3, authorNames=Cui Bingyan, Deng Jia, Zhang Xiang, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Cui Bingyan, Deng Jia, Zhang Xiang. Hand gesture recognition based on surface electromyography[J]. Science Technology and Engineering, 2023, 23(35): 15133-15141., articleTitle=Hand gesture recognition based on surface electromyography, refAbstract=null), Reference(id=1245407879684010117, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2018, volume=18, issue=17, pageStart=61, pageEnd=66, url=null, language=null, rfNumber=[4], rfOrder=4, authorNames=史小华, 卢浩, 廖梓宇, journalName=科学技术与工程, refType=null, unstructuredReference=史小华, 卢浩, 廖梓宇, . 基于表面肌电信号的下肢康复主动训练[J]. 科学技术与工程, 2018, 18(17): 61-66., articleTitle=基于表面肌电信号的下肢康复主动训练, refAbstract=null), Reference(id=1245407879818227854, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2018, volume=18, issue=17, pageStart=61, pageEnd=66, url=null, language=null, rfNumber=[4], rfOrder=5, authorNames=Shi Xiaohua, Lu Hao, Liao Ziyu, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Shi Xiaohua, Lu Hao, Liao Ziyu, et al. Active training of lower limb rehabilitation based on surface electromyography signal[J]. Science Technology and Engineering, 2018, 18(17): 61-66., articleTitle=Active training of lower limb rehabilitation based on surface electromyography signal, refAbstract=null), Reference(id=1245407879939862676, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2023, volume=37, issue=2, pageStart=190, pageEnd=197, url=null, language=null, rfNumber=[5], rfOrder=6, authorNames=黄圣, 茅健, journalName=上海工程技术大学学报, refType=null, unstructuredReference=黄圣, 茅健. 基于改进深度森林的表面肌电手势识别方法[J]. 上海工程技术大学学报, 2023, 37(2): 190-197., articleTitle=基于改进深度森林的表面肌电手势识别方法, refAbstract=null), Reference(id=1245407880057303193, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2023, volume=37, issue=2, pageStart=190, pageEnd=197, url=null, language=null, rfNumber=[5], rfOrder=7, authorNames=Huang Sheng, Mao Jian, journalName=Journal of Shanghai University of Engineering Science, refType=null, unstructuredReference=Huang Sheng, Mao Jian. Recognition method of sEMG gesture based on improved deep forest[J]. Journal of Shanghai University of Engineering Science, 2023, 37(2): 190-197., articleTitle=Recognition method of sEMG gesture based on improved deep forest, refAbstract=null), Reference(id=1245407880208298146, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2017, volume=9, issue=8, pageStart=147, pageEnd=null, url=null, language=null, rfNumber=[6], rfOrder=8, authorNames=Ai Q, journalName=Symmetry, refType=null, unstructuredReference=Ai Q. Research on lower limb motion recognition based on fusion of sEMG and accelerometer signals[J]. Symmetry, 2017, 9(8): 147., articleTitle=Research on lower limb motion recognition based on fusion of sEMG and accelerometer signals, refAbstract=null), Reference(id=1245407880325738662, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2019, volume=19, issue=32, pageStart=219, pageEnd=224, url=null, language=null, rfNumber=[7], rfOrder=9, authorNames=刘国平, 王南星, 周毅, journalName=科学技术与工程, refType=null, unstructuredReference=刘国平, 王南星, 周毅, . 基于改进ReliefF算法的哑铃动作识别[J]. 科学技术与工程, 2019, 19(32): 219-224., articleTitle=基于改进ReliefF算法的哑铃动作识别, refAbstract=null), Reference(id=1245407880418013358, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2019, volume=19, issue=32, pageStart=219, pageEnd=224, url=null, language=null, rfNumber=[7], rfOrder=10, authorNames=Liu Guoping, Wang Nanxing, Zhou Yi, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Liu Guoping, Wang Nanxing, Zhou Yi, et al. Recognition of dumbbell movements based on an improved reliefF algorithm[J]. Science Technology and Engineering, 2019, 19(32): 219-224., articleTitle=Recognition of dumbbell movements based on an improved reliefF algorithm, refAbstract=null), Reference(id=1245407880610951350, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2023, volume=37, issue=6, pageStart=59, pageEnd=67, url=null, language=null, rfNumber=[8], rfOrder=11, authorNames=赵世昊, 周建华, 伏云发, journalName=电子测量与仪器学报, refType=null, unstructuredReference=赵世昊, 周建华, 伏云发. 注意力机制CNN结合肌电特征矩阵的手势识别研究[J]. 电子测量与仪器学报, 2023, 37(6): 59-67., articleTitle=注意力机制CNN结合肌电特征矩阵的手势识别研究, refAbstract=null), Reference(id=1245407880711614651, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2023, volume=37, issue=6, pageStart=59, pageEnd=67, url=null, language=null, rfNumber=[8], rfOrder=12, authorNames=Zhao Shihao, Zhou Jianhua, Fu Yunfa, journalName=Journal of Electronic Measurement and Instrumentation, refType=null, unstructuredReference=Zhao Shihao, Zhou Jianhua, Fu Yunfa. Investigation of gesture recognition using attention mechanism CNN combined electromyography feature matrix[J]. Journal of Electronic Measurement and Instrumentation, 2023, 37(6): 59-67., articleTitle=Investigation of gesture recognition using attention mechanism CNN combined electromyography feature matrix, refAbstract=null), Reference(id=1245407880803889346, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2019, volume=50, issue=3, pageStart=2265, pageEnd=2280, url=null, language=null, rfNumber=[9], rfOrder=13, authorNames=Xi X G, Yang C, Shi J H, journalName=Neural Processing Letters, refType=null, unstructuredReference=Xi X G, Yang C, Shi J H, et al. Surface electromyography-based daily activity recognition using wavelet coherence coefficient and support vector machine[J]. Neural Processing Letters, 2019, 50(3): 2265-2280., articleTitle=Surface electromyography-based daily activity recognition using wavelet coherence coefficient and support vector machine, refAbstract=null), Reference(id=1245407880912941257, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2022, volume=22, issue=19, pageStart=8370, pageEnd=8377, url=null, language=null, rfNumber=[10], rfOrder=14, authorNames=刘光达, 许蓝予, 肖若兰, journalName=科学技术与工程, refType=null, unstructuredReference=刘光达, 许蓝予, 肖若兰, . 基于FGNN算法的sEMG肌肉疲劳分类方法[J]. 科学技术与工程, 2022, 22(19): 8370-8377., articleTitle=基于FGNN算法的sEMG肌肉疲劳分类方法, refAbstract=null), Reference(id=1245407881005215950, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2022, volume=22, issue=19, pageStart=8370, pageEnd=8377, url=null, language=null, rfNumber=[10], rfOrder=15, authorNames=Liu Guangda, Xu Lanyu, Xiao Ruolan, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Liu Guangda, Xu Lanyu, Xiao Ruolan, et al. Muscle fatigue classification method of sEMG signal based on FGNN algorithm[J]. Science Technology and Engineering, 2022, 22(19): 8370-8377., articleTitle=Muscle fatigue classification method of sEMG signal based on FGNN algorithm, refAbstract=null), Reference(id=1245407881105879249, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2019, volume=176, issue=null, pageStart=228, pageEnd=235, url=null, language=null, rfNumber=[11], rfOrder=16, authorNames=Sui X W, Wan K X, Zhang Y, journalName=Optik, refType=null, unstructuredReference=Sui X W, Wan K X, Zhang Y. Pattern recognition of sEMG based on wavelet packet transform and improved SVM[J]. Optik, 2019, 176: 228-235., articleTitle=Pattern recognition of sEMG based on wavelet packet transform and improved SVM, refAbstract=null), Reference(id=1245407881185571032, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2022, volume=34, issue=6, pageStart=105, pageEnd=112, url=null, language=null, rfNumber=[12], rfOrder=17, authorNames=刘声中, 许德章, journalName=嘉兴学院学报, refType=null, unstructuredReference=刘声中, 许德章. 一种基于改进sEMG活动段检测的手部运动意图识别新方法[J]. 嘉兴学院学报, 2022, 34(6): 105-112., articleTitle=一种基于改进sEMG活动段检测的手部运动意图识别新方法, refAbstract=null), Reference(id=1245407881328177369, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2022, volume=34, issue=6, pageStart=105, pageEnd=112, url=null, language=null, rfNumber=[12], rfOrder=18, authorNames=Liu Shengzhong, Xu Dezhang, journalName=Journal of Jiaxing University, refType=null, unstructuredReference=Liu Shengzhong, Xu Dezhang. A new method for hand motion intent recognition based on improved sEMG signal active segment detection[J]. Journal of Jiaxing University, 2022, 34(6): 105-112., articleTitle=A new method for hand motion intent recognition based on improved sEMG signal active segment detection, refAbstract=null), Reference(id=1245407881479172321, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2018, volume=43, issue=12, pageStart=7817, pageEnd=7835, url=null, language=null, rfNumber=[13], rfOrder=19, authorNames=Gupta R, Agarwal R, journalName=Arabian Journal for Science and Engineering, refType=null, unstructuredReference=Gupta R, Agarwal R. Electromyographic signal-driven continuous locomotion mode identification module design for lower limb prosthesis control[J]. Arabian Journal for Science and Engineering, 2018, 43(12): 7817-7835., articleTitle=Electromyographic signal-driven continuous locomotion mode identification module design for lower limb prosthesis control, refAbstract=null), Reference(id=1245407881609195753, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2006, volume=8, issue=1, pageStart=11, pageEnd=35, url=null, language=null, rfNumber=[14], rfOrder=20, authorNames=Reaz M, Hussain M, Mohd-Yasin F, journalName=Biological Procedures Online, refType=null, unstructuredReference=Reaz M, Hussain M, Mohd-Yasin F. Techniques of EMG signal analysis: detection, processing, classification and applications (Correction)[J]. Biological Procedures Online, 2006, 8(1): 11-35., articleTitle=Techniques of EMG signal analysis: detection, processing, classification and applications (Correction), refAbstract=null), Reference(id=1245407881747607790, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2021, volume=21, issue=12, pageStart=13019, pageEnd=13028, url=null, language=null, rfNumber=[15], rfOrder=21, authorNames=Xu H, Xiong A, journalName=IEEE Sensors Journal, refType=null, unstructuredReference=Xu H, Xiong A. Advances and disturbances in sEMG-based intentions and movements recognition: a review[J]. IEEE Sensors Journal, 2021, 21(12): 13019-13028., articleTitle=Advances and disturbances in sEMG-based intentions and movements recognition: a review, refAbstract=null), Reference(id=1245407881839882484, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2008, volume=56, issue=8, pageStart=4069, pageEnd=4081, url=null, language=null, rfNumber=[16], rfOrder=22, authorNames=Chiang J, Wang Z J, McKeown M, journalName=IEEE Transactions on Signal Processing, refType=null, unstructuredReference=Chiang J, Wang Z J, & McKeown M. A hidden Markov, multivariate autoregressive (HMM-mAR) network framework for analysis of surface EMG (sEMG) data[J]. IEEE Transactions on Signal Processing, 2008, 56(8): 4069-4081., articleTitle=A hidden Markov, multivariate autoregressive (HMM-mAR) network framework for analysis of surface EMG (sEMG) data, refAbstract=null), Reference(id=1245407881957322998, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2024, volume=50, issue=2, pageStart=9, pageEnd=15, url=null, language=null, rfNumber=[17], rfOrder=23, authorNames=王睿, 杨建荣, journalName=四川建筑科学研究, refType=null, unstructuredReference=王睿, 杨建荣. 基于长短时记忆神经网络易损性分析的适用性研究[J]. 四川建筑科学研究, 2024, 50(2): 9-15., articleTitle=基于长短时记忆神经网络易损性分析的适用性研究, refAbstract=null), Reference(id=1245407882104123646, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, doi=null, pmid=null, pmcid=null, year=2024, volume=50, issue=2, pageStart=9, pageEnd=15, url=null, language=null, rfNumber=[17], rfOrder=24, authorNames=Wang Rui, Yang Jianrong, journalName=Sichuan Building Science, refType=null, unstructuredReference=Wang Rui, Yang Jianrong. Study on the applicability of vulnerability analysis based on long short-term memory neural network[J]. Sichuan Building Science, 2024, 50(2): 9-15., articleTitle=Study on the applicability of vulnerability analysis based on long short-term memory neural network, refAbstract=null)], funds=[Fund(id=1245407879021310036, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, awardId=52365039, language=CN, fundingSource=国家自然科学基金(52365039), fundOrder=null, country=null), Fund(id=1245407879113584734, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, awardId=2023TSYCLJ0051, language=CN, fundingSource=自治区天山英才项目(2023TSYCLJ0051), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1245407871408648552, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, xref=1, ext=[AuthorCompanyExt(id=1245407871417037162, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China), AuthorCompanyExt(id=1245407871421231467, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871408648552, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017)]), AuthorCompany(id=1245407871568032125, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, xref=2, ext=[AuthorCompanyExt(id=1245407871580615040, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871568032125, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 Medical Laboratory Center, The First Affiliated Hospital of Xinjiang Medical University, Urumqi 830054, China), AuthorCompanyExt(id=1245407871597392257, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, companyId=1245407871568032125, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 新疆医科大学第一附属医院, 医学检验中心, 乌鲁木齐 830054)])], figs=[ArticleFig(id=1245407875472929588, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Fig.1, caption=Diagram of the experimental scene, figureFileSmall=evwqks+BSKx3HMDm4pvg+A==, figureFileBig=AMoDOmZkfPbW8k41X2R+yA==, tableContent=null), ArticleFig(id=1245407875569398588, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=图1, caption=实验场景图, figureFileSmall=evwqks+BSKx3HMDm4pvg+A==, figureFileBig=AMoDOmZkfPbW8k41X2R+yA==, tableContent=null), ArticleFig(id=1245407875875582806, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Fig.2, caption=Sliding window segmentation diagram, figureFileSmall=BmoeRIg+2DshnbDzGWoooQ==, figureFileBig=vN1FXpUsBy6TpYBxwfUGtw==, tableContent=null), ArticleFig(id=1245407875993023330, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=图2, caption=滑动窗口分割图, figureFileSmall=BmoeRIg+2DshnbDzGWoooQ==, figureFileBig=vN1FXpUsBy6TpYBxwfUGtw==, tableContent=null), ArticleFig(id=1245407876097880941, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Fig.3, caption=1DCNN model structure, figureFileSmall=54qeWm2hySg+qrOQDEb5zQ==, figureFileBig=yc1AjP1QB3Jz0TjfOxNq+g==, tableContent=null), ArticleFig(id=1245407876211127159, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=图3, caption=1DCNN模型结构, figureFileSmall=54qeWm2hySg+qrOQDEb5zQ==, figureFileBig=yc1AjP1QB3Jz0TjfOxNq+g==, tableContent=null), ArticleFig(id=1245407876353733504, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Fig.4, caption=LSTM model structure, figureFileSmall=J5PpBYXK/zLUH/q0KpciFQ==, figureFileBig=9fpT0q4xEZNZdV2rcgCAdg==, tableContent=null), ArticleFig(id=1245407876492145547, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=图4, caption=LSTM模型结构, figureFileSmall=J5PpBYXK/zLUH/q0KpciFQ==, figureFileBig=9fpT0q4xEZNZdV2rcgCAdg==, tableContent=null), ArticleFig(id=1245407876680889241, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Fig.5, caption=CNN-LSTM overall architecture, figureFileSmall=4gH5bE9c7K5cFh7VSHDw4Q==, figureFileBig=VGChKfXLc8j7lUrFIZyiWg==, tableContent=null), ArticleFig(id=1245407876840272804, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=图5, caption=CNN-LSTM总体架构, figureFileSmall=4gH5bE9c7K5cFh7VSHDw4Q==, figureFileBig=VGChKfXLc8j7lUrFIZyiWg==, tableContent=null), ArticleFig(id=1245407876970296241, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Fig.6, caption=The loss curve of the CNN-LSTM model, figureFileSmall=FgjCF2Ysf253Qi1sK3HG3Q==, figureFileBig=HbB/PhSeveJaP5rS8v67hg==, tableContent=null), ArticleFig(id=1245407877087736764, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=图6, caption=CNN-LSTM模型损失曲线, figureFileSmall=FgjCF2Ysf253Qi1sK3HG3Q==, figureFileBig=HbB/PhSeveJaP5rS8v67hg==, tableContent=null), ArticleFig(id=1245407877192594371, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Fig.7, caption=Comparison of the performance of the 4 classification models, figureFileSmall=cO2MvTGV1msJfpwilsGCsg==, figureFileBig=jq/YqIr8CaT0mBn0U9AQaQ==, tableContent=null), ArticleFig(id=1245407877347783630, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=图7, caption=4种分类模型性能对比, figureFileSmall=cO2MvTGV1msJfpwilsGCsg==, figureFileBig=jq/YqIr8CaT0mBn0U9AQaQ==, tableContent=null), ArticleFig(id=1245407877482001365, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Fig.8, caption=ROC curves for 4 classification models, figureFileSmall=+xYRh8UFVWgO8D7e8xY9Vw==, figureFileBig=lbnNLb4bPshW+2Vy5CnVXQ==, tableContent=null), ArticleFig(id=1245407877603636193, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=图8, caption=4种分类模型ROC曲线, figureFileSmall=+xYRh8UFVWgO8D7e8xY9Vw==, figureFileBig=lbnNLb4bPshW+2Vy5CnVXQ==, tableContent=null), ArticleFig(id=1245407877725271020, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Fig.9, caption=Confusion matrices for 4 models, figureFileSmall=nw7+SYe626OdiL9qpCn3zQ==, figureFileBig=TouK0kZTXKijNwJE0385kA==, tableContent=null), ArticleFig(id=1245407877855294455, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=图9, caption=4种模型的混淆矩阵, figureFileSmall=nw7+SYe626OdiL9qpCn3zQ==, figureFileBig=TouK0kZTXKijNwJE0385kA==, tableContent=null), ArticleFig(id=1245407878018872319, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Table 1, caption=

Basic information of the subject

, figureFileSmall=null, figureFileBig=null, tableContent=
年龄/岁 身高/cm 体重/kg
22~28 160~189 60~85
), ArticleFig(id=1245407878123728901, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=表1, caption=

受试者基本信息

, figureFileSmall=null, figureFileBig=null, tableContent=
年龄/岁 身高/cm 体重/kg
22~28 160~189 60~85
), ArticleFig(id=1245407878236975121, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Table 2, caption=

Feature calculation formula

, figureFileSmall=null, figureFileBig=null, tableContent=
特征 计算公式
时域 MAV TMAV=1/N$\sum _{i=1}^{N}$|xi|
RMS TRMS=$\sqrt{1/N\sum _{i=1}^{N}{x}_{i}^{2}}$
频域 MPF FMPF=${\int }_{0}^{+\infty } $fP(f)df/${\int }_{0}^{+\infty } $P(f)df
MF ${\int }_{0}^{{F}_{MF}} $P(f)df=$\frac{1}{2}{\int }_{0}^{+\infty } $P(f)df
), ArticleFig(id=1245407878383775775, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=表2, caption=

特征计算公式

, figureFileSmall=null, figureFileBig=null, tableContent=
特征 计算公式
时域 MAV TMAV=1/N$\sum _{i=1}^{N}$|xi|
RMS TRMS=$\sqrt{1/N\sum _{i=1}^{N}{x}_{i}^{2}}$
频域 MPF FMPF=${\int }_{0}^{+\infty } $fP(f)df/${\int }_{0}^{+\infty } $P(f)df
MF ${\int }_{0}^{{F}_{MF}} $P(f)df=$\frac{1}{2}{\int }_{0}^{+\infty } $P(f)df
), ArticleFig(id=1245407878492827690, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=null, caption=null, figureFileSmall=null, figureFileBig=null, tableContent=
网络层 卷积核
数目
卷积核
尺寸
步长 输出
维度
卷积层1 64 3×1 5 64×1 020
BN1+ReLU 64 64×1 020
池化层1 64 2×2 64×510
卷积层2 32 3×1 3 32×508
BN2+ReLU 32 32×508
池化层2 32 2×2 32×254
卷积层3 10 3×1 3 10×252
BN3+ReLU 10 10×252
池化层3 10 2×2 10×126
全连接层1 1 1 260 1 260×1
全连接层2 1 600 600×1
全连接层3 1 100 100×1
), ArticleFig(id=1245407878606073908, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=表3, caption=

1DCNN模型参数配置 Table 3 1DCNN model parameter configuration

, figureFileSmall=null, figureFileBig=null, tableContent=
网络层 卷积核
数目
卷积核
尺寸
步长 输出
维度
卷积层1 64 3×1 5 64×1 020
BN1+ReLU 64 64×1 020
池化层1 64 2×2 64×510
卷积层2 32 3×1 3 32×508
BN2+ReLU 32 32×508
池化层2 32 2×2 32×254
卷积层3 10 3×1 3 10×252
BN3+ReLU 10 10×252
池化层3 10 2×2 10×126
全连接层1 1 1 260 1 260×1
全连接层2 1 600 600×1
全连接层3 1 100 100×1
), ArticleFig(id=1245407878715125821, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=EN, label=Table 4, caption=

LSTM model parameters configuration

, figureFileSmall=null, figureFileBig=null, tableContent=
参数名称 符号或数值
损失函数 Cross Entropy
优化器 Adadelta
层数 3
LSTM单元 100
批量大小 40
学习率 0.001
), ArticleFig(id=1245407878845149252, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1245407864299303881, language=CN, label=表4, caption=

LSTM模型参数配置

, figureFileSmall=null, figureFileBig=null, tableContent=
参数名称 符号或数值
损失函数 Cross Entropy
优化器 Adadelta
层数 3
LSTM单元 100
批量大小 40
学习率 0.001
)], attaches=null, journal=Journal(id=1146119176004939786, delFlag=0, nameCn=科学技术与工程, nameEn=Science Technology and Engineering, nameHistory1=null, nameHistory2=null, issn=1671-1815, eissn=, cn=11-4688/T, coden=null, periodic=4, language=CN, oaType=是, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=UKU/O7GSka5polgCTkbIIw==, journalPrice=null, startedYear=null, abbrevIsoEn=Sci Technol Eng, journalRemark=null, publicationField=null, createdTime=null, updatedTime=1754445529766, createdBy=null, updatedBy=13701087609, firstLetterCn=S, firstLetterEn=S, subjectCode=Natural Sciences, subjectName=自然科学, subjectCodeEn=Natural Sciences, subjectNameEn=null, picCn=UKU/O7GSka5polgCTkbIIw==, picEn=5hwlULoNwcbj3xUmVi9MAQ==, jcr=null, cjcr=null, exts=[JournalExt(id=1159791870395564357, language=CN, name=科学技术与工程, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=null, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=http://www.stae.com.cn/jsygc/home, createdTime=1754445529793, updatedTime=1754445529793, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=http://www.stae.com.cn/jsygc/site/menus/20090429150146001, submissionAuthorUrl=http://www.stae.com.cn/jsygc/author/login, submissionEditorUrl=http://www.stae.com.cn/jsygc/editor/login, submissionReviewUrl=http://www.stae.com.cn/jsygc/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1159791870441701702, language=EN, name=Science Technology and Engineering, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=null, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=http://www.stae.com.cn/jsygc/home, createdTime=1754445529804, updatedTime=1754445529804, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=http://www.stae.com.cn/jsygc/author/login, submissionEditorUrl=http://www.stae.com.cn/jsygc/editor/login, submissionReviewUrl=http://www.stae.com.cn/jsygc/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1146123166801305609, websiteList=[Website(id=1148243202391400884, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146123166801305609, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/kxjsygc/CN, language=CN, createTime=1751692112777, createBy=18614031015, updateTime=1753520965431, updateBy=18614031015, name=科学技术与工程-中文站点, tplId=1146099689490845704, title=科学技术与工程, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1148622798802673703, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=articleTextType, value=kx, createTime=1751782615614, updateTime=1751782615614, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798781702180, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=banner, value=null, createTime=1751782615609, updateTime=1751782615609, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798769119267, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=j86gbwi+p0Idkyl5SzIlmQ==, createTime=1751782615606, updateTime=1751782615606, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798794285094, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1751782615612, updateTime=1751782615612, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798790090789, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1751782615611, updateTime=1751782615611, creator=18614031015, updator=18614031015)]), Website(id=1155914124811976731, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146123166801305609, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/kxjsygc/EN, language=EN, createTime=1753521003206, createBy=18614031015, updateTime=1753521003206, updateBy=18614031015, name=科学技术与工程-英文站点, tplId=1146101810881728533, title=Science Technology and Engineering, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1155914371227308235, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=articleTextType, value=kx, createTime=1753521061952, updateTime=1753521061952, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371210531016, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=banner, value=null, createTime=1753521061947, updateTime=1753521061947, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371202142407, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=j86gbwi+p0Idkyl5SzIlmQ==, createTime=1753521061945, updateTime=1753521061945, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371223113930, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1753521061950, updateTime=1753521061950, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371218919625, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1753521061949, updateTime=1753521061949, creator=18614031015, updator=18614031015)])], journalTitle=科学技术与工程, weixinUrl=null, journalUrl=null, iacademicId=null, status=0, seqNo=null, journalTitleEn=Science Technology and Engineering, journalPhotoCn=UKU/O7GSka5polgCTkbIIw==, journalPhotoEn=5hwlULoNwcbj3xUmVi9MAQ==, journalFirstLetter=S, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=null, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/kxjsygc/CN/10.12404/j.issn.1671-1815.2403062, detailUrlEn=https://castjournals.cast.org.cn/joweb/kxjsygc/EN/10.12404/j.issn.1671-1815.2403062, pdfUrlCn=https://castjournals.cast.org.cn/joweb/kxjsygc/CN/PDF/10.12404/j.issn.1671-1815.2403062, pdfUrlEn=https://castjournals.cast.org.cn/joweb/kxjsygc/EN/PDF/10.12404/j.issn.1671-1815.2403062, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于表面肌电信号的CNN-LSTM模型下肢动作识别
收藏切换
PDF下载
周智伟 1 , 陶庆 1, * , 苏娜 2 , 刘景轩 1 , 李博文 1 , 裴浩 1
科学技术与工程 | 论文·自动化技术、计算机技术 2025,25(7): 2841-2848
收起
收藏切换
科学技术与工程 | 论文·自动化技术、计算机技术 2025, 25(7): 2841-2848
基于表面肌电信号的CNN-LSTM模型下肢动作识别
全屏
周智伟1 , 陶庆1, * , 苏娜2, 刘景轩1, 李博文1, 裴浩1
作者信息
  • 1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017
  • 2 新疆医科大学第一附属医院, 医学检验中心, 乌鲁木齐 830054
  • 周智伟(1999—),男,汉族,新疆奎屯人,硕士研究生。研究方向:下肢动作识别、深度学习。E-mail:

通讯作者:

* 陶庆(1978—),男,汉族,江苏常熟人,博士,教授。研究方向:先进设计与制造、虚拟现实技术。E-mail:
Lower Limb Motion Recognition Based on Surface Electromyography and a CNN-LSTM Fusion Model
Zhi-wei ZHOU1 , Qing TAO1, * , Na SU2, Jing-xuan LIU1, Bo-wen LI1, Hao PEI1
Affiliations
  • 1 College of Intelligent Manufacturing Modern Industry, Xinjiang University, Urumqi 830017, China
  • 2 Medical Laboratory Center, The First Affiliated Hospital of Xinjiang Medical University, Urumqi 830054, China
出版时间: 2025-03-08 doi: 10.12404/j.issn.1671-1815.2403062
文章导航
收藏切换

为了提高对下肢运动的分类准确度,提出了一种基于表面肌电信号(surface electromyography, sEMG)的卷积神经网络与长短期记忆网络融合识别模型(convolutional neural network and long short-term memory network, CNN-LSTM)。首先,采集了20名受试者进行上楼、下楼、行走和蹲起4种步态动作的sEMG;接着,对采集到的sEMG数据进行预处理,并提取了两种时域和频域特征,用作机器学习识别模型的特征输入;最后,基于预处理后肌电信号数据,构建了CNN-LSTM的下肢动作识别模型,并与CNN、LSTM和支持向量机(support vector machine, SVM)模型的性能进行对比。结果显示,CNN-LSTM模型在下肢动作识别准确率上分别比CNN、LSTM和SVM模型高出2.16%、8.34%、和11.16%,证明了其优越的分类性能。研究结论为康复医疗器械与助力器械提供了一个有效的下肢运动功能改善方案。

表面肌电信号  /  下肢动作识别  /  CNN-LSTM  /  卷积神经网络  /  长短时记忆网络

To enhance the classification accuracy of lower limb movements, this paper was introduced a hybrid recognition model based on surface electromyography (sEMG) that combines convolutional neural networks (CNN) with long short-term memory networks (LSTM). Initially, sEMG data were collected from 20 subjects performing four types of gait movements: ascending stairs, descending stairs, walking, and squatting. Subsequently, the collected sEMG data underwent preprocessing, and both time domain and frequency domain features were extracted to serve as inputs for the machine learning recognition model. The CNN-LSTM model was then constructed for lower limb action recognition and compared against the performances of CNN, LSTM, and SVM (support vector machine,)models. The results demonstrate that the CNN-LSTM model outperforms the CNN, LSTM, and SVM models by 2.16%, 8.34%, and 11.16% in accuracy, respectively, thereby proving its superior classification performance. This model provides an effective solution for enhancing lower limb motor functions, offering significant benefits for rehabilitation medical equipment and power assist devices.

surface electromyographic signals  /  lower limb motion recognition  /  CNN-LSTM  /  convolutional neural networks  /  long short-term memory networks
周智伟, 陶庆, 苏娜, 刘景轩, 李博文, 裴浩. 基于表面肌电信号的CNN-LSTM模型下肢动作识别. 科学技术与工程, 2025 , 25 (7) : 2841 -2848 . DOI: 10.12404/j.issn.1671-1815.2403062
Zhi-wei ZHOU, Qing TAO, Na SU, Jing-xuan LIU, Bo-wen LI, Hao PEI. Lower Limb Motion Recognition Based on Surface Electromyography and a CNN-LSTM Fusion Model[J]. Science Technology and Engineering, 2025 , 25 (7) : 2841 -2848 . DOI: 10.12404/j.issn.1671-1815.2403062
下肢运动的识别在众多领域中具有显著的应用价值。在康复治疗领域,下肢运动识别可以帮助医生客观地评估患者的身体状态,从而优化康复方案。此外,下肢运动识别使人体与可穿戴动力辅助外骨骼机器人有效结合,从而提高人体的负重能力和工作效率。据报道,识别下肢运动也有望改善老年、体弱和运动障碍患者的日常生活质量。表面肌电信号(surface electromyography, sEMG)由多个运动单元产生的动作电位叠加而成,其中蕴含了与人类行为密切相关的生物学信息。相较于肌内肌电信号的采集,sEMG的采集过程具有安全性高、可靠性强和无创性等显著优势[1]。表面肌电信号能够直接反映肌肉的活动模式,在康复机器人、可穿戴外骨骼控制以及人机交互中得到了广泛的应用。
在表面肌电信号和动态脑电信号的预处理研究方面,主要集中在特征的提取方法与分类算法,旨在实现对多种人体肢体运动模式的高识别率。例如,Totty等[2]采用了K-最近邻(K-nearest neighbor,KNN)算法结合肌肉激活数据和运动数据,对功能性手臂活动行为观察系统中的日常生活活动进行了有效分类。崔冰艳等[3]结合时域与时频域特征的sEMG特征空间构造方法,提出了基于sEMG的支持向量机(support vector machine, SVM)分类器、人工鱼群优化算法(artificial fish swarm algorithm, AFSA) 优化的SVM分类器和深度学习卷积神经网络(convolutional neural network,CNN)分类模型,对上肢手势动作进行了识别。史小华等[4]提出了一种基于表面肌电信号的下肢康复训练模式,通过提取肌电信号特征,并利用反向传播神经网络(back-propagation neural network, BPNN)辨识运动意图,将结果作为康复机器人驱动信号,实现主动训练。黄圣等[5]提出一种使用sEMG数据进行手部动作识别的改进的深度森林模型,利用3类分类器算法之间的互补性,提高深度森林模型的分类识别准确率,对16种常用手部动作的平均识别精度为94%。Ai等[6]从表面肌电信号中提取时域特征和小波系数,并对加速度信号使用动态时间规划(dynamic time warping,DTW)距离进行特征提取,采用线性判别分析(linear discriminant analysis,LDA)和SVM对5种下肢动作进行分类。刘国平等[7]通过改进的ReliefF算法对数据预处理后的时域特征向量进行最优选择,并使用基于平衡决策树的支持向量机对5种常见的哑铃动作进行分层识别,为个性化哑铃动作指导奠定了基础。赵世昊等[8]采用数据增强后的EFM(enhanced feature map)作为输入样本,并在CNN架构中融入ECA(efficient channel attention)机制以赋予关键特征信息更高的权重,进一步通过模块化设计调整网络中深层特征提取模块的层数,有效地提高了手势识别的准确度。Xi等[9]应用小波变换处理表面肌电信号,并计算得到小波相干系数,使用SVM对6种日常活动进行分类。刘光达等[10]提出了一种基于肌电信号的肌肉疲劳分类方法,将遗传算法的交叉变异和改进的果蝇优化算法混合,并与神经网络结合对肌肉疲劳进行识别,实现了肌肉疲劳的准确检测和分类。Sui等[11]通过运用小波包变换(wavelet packet transform,WPT)来分解表面肌电信号,以提取小波包系数,并从中计算得到的方差和能量作为输入,喂入改进的SVM分类器中,对六种上肢动作的识别实现了90.66%的准确率。刘声中等[12]通过改进sEMG活动段检测技术,提高了手部运动意图分类的精确度。采用了双级判别自适应门限技术检测sEMG的活动段,并将运动意图对应的特征矩阵及其标签用作LSTM手部运动意图分类模型的输入与输出,使得6种运动意图平均分类精度提升至91.7%。Gupta等[13]记录了18名受试者在3种下肢运动模式下的双通道肌电信号,并评估了窗口大小、特征向量类型及分类器类型对识别性能的影响,发现选用256 ms窗口大小、32 ms重叠、LDA分类器以及时间特征向量可获得优良性能。
根据文献[14-15]的研究发现,当前的运动识别技术主要依赖于肌电信号和脑电信号,其中使用表面肌电信号进行下肢动作识别能够在客观条件下实现满意的识别成果。但是,现有的研究下肢动作分类方法一般先提取sEMG的特征,然后使用机器学习或深度学习算法对所提取的特征进行识别分类,以实现对下肢动作的识别。然而,这种手动提取特征方法只考虑了部分信息,不能完全反映输入数据的整体特性,无法提取出关键的特征,导致在特征提取过程中部分sEMG信息缺失,从而影响识别率。因此,为解决上述传统特征提取过程中的信息丢失的问题,本研究基于表面肌电信号提出了CNN-LSTM的下肢运动识别模型,该模型能够有效地捕捉肌电信号中的时空特征,提高了下肢动作的识别精度和鲁棒性,能够为下肢运动康复提供更为可靠和高效的解决方案。
采集20名的健康受试者的sEMG,其各项身体基本参数如表1所示。
在实验开始前,所有受试者均没有遭受骨折、扭伤、肌肉拉伤或其他可能影响运动功能的伤害。且在实验前一周无任何剧烈运动,有效避免了肌肉酸痛或不适。实验设备选用PLUX无线肌电采集设备,选择1 000 Hz的采样频率。此外,选用氯化银(AgCl)材质的肌电电极片和75%酒精棉。
图1所示,受试者分别进行了4种下肢运动:上楼、蹲起、行走和下楼。其中,每名受试者在进行上楼和下楼这两种下肢运动模式时,每个动作进行8组,每组需进行5次,每次运动周期为2 s;进行蹲起运动时需进行3组,每组进行20次,每次运动周期为3 s;进行行走运动需进行5组,每组运动周期为60 s,可得到30组数据,组间休息3~5 min,并且在不同的动作之间需要休息10~15 min,以避免肌肉疲劳影响结果。
在sEMG采集过程中,易受环境噪声和其他生物电信号(如心电信号)干扰的特性,故需要进行降噪处理。sEMG中包含有用信息的频率范围主要集中在5~200 Hz。因此,本研究采用了30~300 Hz的四阶巴特沃斯带通滤波器与50 Hz的陷波滤波器来对原始sEMG进行滤波。此外,利用db2小波基进行4层小波分解,并通过小波阈值技术进一步实现降噪。随后,根据各动作周期的初始点将处理后的数据进行分段,其中行走、上楼梯和下楼梯的动作周期定为2 s,而蹲起动作的周期设定为3 s。采用重叠窗技术对每个动作周期内的sEMG数据段进行进一步的分割。这里选用了1 024 ms的滑动窗口配合512 ms的步进长度,以此将各通道的sEMG细分成序列化的数据窗口。图2为使用滑动窗口技术对单个通道的肌电信号进行分割的过程。
为了从处理后的肌电数据中提取有效信息,供机器学习模型作为特征输入,本研究从采集的肌电信号中同时提取了时域和频域特征,并对这些特征进行了归一化处理。在时域分析中,关键指标包括肌电信号时间序列xi的绝对平均值(mean absolute value, MAV)和均方根值(root mean square, RMS)。在频域分析中,通过分析肌电信号的功率谱密度函数P(f),进一步提取平均功率频率(mean power frequency, MPF)和中值频率(median frequency, MF)作为重要的频域特征,计算公式如表2所示。
sEMG被认为是一种混合信号,其源于多个肌肉活动的时间和空间上的叠加,这导致了其较高的复杂性[16]。传统的卷积神经网络多用于图像的处理,难以有效地捕捉一维时间序列数据中的特征。相比之下,一维卷积神经网络(1DCNN)可以更直接地处理时序数据,使模型的参数量更少,并且能够从这些时序顺序中学习特征。因此,本文中选择1DCNN作为处理sEMG的模型,为动作识别提供有力支持。模型结构如图3所示。
使用包含3个卷积层、池化层、批量标准化层(batch normalization, BN)、3个全连接层以及一个Dropout层的1DCNN模型。在卷积层中,输入信号会经过一维卷积处理,计算公式为
$h_{j}^{l}=\sum_{i=1}^{n} x_{i}^{l} \otimes k_{i j}^{l}+b_{j}^{l}$
式(1)中:l为层数;j为元素序数;n为输入特征向量的长度;h为输出特征向量;x为输入特征向量;k为卷积核;b为偏置向量;􀱋为卷积操作。
为了增强模型的训练效率与稳定性,在每个卷积层后使用BN归一化,计算过程如下。
首先,沿Batch的维度计算相同通道的均值与方差,计算公式为
μB=$\frac{1}{n}\sum _{i=1}^{n}$xi
${\sigma }_{B}^{2}$=$\frac{1}{n}\sum _{i=1}^{n}({x}_{i}-{\mu }_{B}{)}^{2}$
然后,对相同Batch维度上的均值和方差做归一化,计算公式为
${\stackrel{\wedge }{x}}_{i}$=$\frac{{x}_{i}-{\mu }_{B}}{\sqrt{{\sigma }_{B}^{2}+\epsilon }}$
式(4)中: ε为一个较小的正数,通常用于确保数值稳定性,避免分母出现零值的情况。
其次,通过学习两个参数γ(缩放因子)和β(平移因子)来恢复特征值可能需要的任何缩放和平移,计算公式为
yi=γ${\stackrel{\wedge }{x}}_{i}$+β
最后,在经过每个BN后采用ReLU激活函数,计算公式为
ReLU(x)=max(0,x)=$\left\{\begin{array}{ll}x,& x\ge 0\\ 0,& x<0\end{array}\right.$
为了提高模型训练速度,本研究采用最大池化进行降采样。在训练阶段,为了评估模型输出的概率分布与真实标签的概率分布之间的偏差,本文中选用交叉熵损失函数,计算公式为
L(θ)=-$\frac{1}{n}\left[\sum _{i=1}^{n}\sum _{k=1}^{K}I\{{y}_{i}=k\}lg\frac{exp\left({{\theta }^{T}}_{K}x\right)}{\sum _{j=1}^{K}exp\left({{\theta }^{T}}_{j}x\right)}\right]$
式(7)中: I {y= k}为一个指示函数,表示当样本i 的真实类别yi等于类别k时,取值为1,否则取值为0;K表示类别数;θ为模型的参数。
此外,为了优化学习过程并最小化损失函数,本研究采用Adadelta优化器来自动调整学习率。在3个全连接层中间插入两个Dropout层,以增强模型的泛化能力并避免过拟合。如表3所示为网络模型配置。
长短时记忆网络(long short-term memory, LSTM)是一种特殊的循环神经网络,通过引入门控单元来控制数据流,从而有效解决传统循环神经网络在处理长序列数据时遭遇的梯度消失和梯度爆炸问题[17]。肌电信号通常包含复杂的时间动态,这些动态可能在较长的时间尺度上展现出相关性,而LSTM的特有结构使其能够学习这些长期依赖关系。LSTM网络的结构如图4所示。
LSTM的工作机制包括:当前输入xt和先前时间步的隐藏状态ht-1通过sigmoid函数处理生成遗忘门输出ft,该输出在0~1之间变化。输出值为0时表示“完全遗忘”,为1时表示“完全保留”。计算公式为
ft=σ(Wf[ht-1,xt]+bf)
式(8)中:wf为权重;σ为sigmoid激活函数;bf为偏置。
输入门在单元状态更新的过程中扮演者关键角色。首先,通过sigmoid确定需要更新的信息部分,生成输入门的输出it。其次,通过tanh层生成一个新的候选值${\stackrel{~}{C}}_{t}$,该向量表示可能被添加到状态中的新信息。最终,将遗忘门的输出ft与前一单元状态Ct-1相乘,以舍弃无用信息,并将输入门的输出it与候选值${\stackrel{~}{C}}_{t}$相乘,以此来更新新的单元状态Ct。计算公式为
it=σ(Wi[ht-1,xt]+bi)
$\tilde{C}_{t}=\tanh \left(W_{c}\left[h_{t-1}, x_{t}\right]+b_{c}\right)$
$C_{t}=f_{t} * C_{t-1}+i_{t} * \tilde{C}_{t}$
式中:$\tilde{C}_{t}$表示候选值;ft*Ct-1表示选择性忘记不相关信息;it*$\tilde{C}_{t}$表示保留有用信息。
在LSTM中,输出门经过sigmoid层决定应被输出的单元状态,将单元状态通过tanh层调整后与sigmoid层的输出相乘,以产生最终的输出ht。计算公式为
Ot=(WO[ht-1,xt]+bo)
ht=Ot*tanh(Ct)
式中:Ot表示输出门的激活函数。
在本研究中LSTM网络参数配置如表4所示。
在上述分析的基础上,本研究构建了一个基于sEMG的CNN-LSTM下肢运动识别模型,其整体框架结构如图5所示。首先,将预处理过的sEMG输入至CNN层进行特征提取;随后经过两个全连接层,然后进入LSTM层;最终,再通过全连接层来生成最终的分类决策。针对多通道EMG信号,CNN能够解析不同肌肉之间的相互作用及其空间布局;而LSTM则优秀地处理了时间序列数据,记忆了信息的长期依赖。通过并联这两种模型,能够对sEMG进行高效且精确的识别。
在对分类任务的性能进行评估中,本论文采用了四项核心统计度量以作为衡量分类模型的评价标准,具体为准确率A(accuracy)、精确率P(precision)、召回率R(recall)及F1(F1-score)。这一组指标综合反映了分类模型性能的多个维度,为模型表现的全面评价提供了依据。计算公式为
A=$\frac{TP+TN}{TP+TN+FP+FN}$×100%
P=$\frac{TP}{TP+FP}$×100%
R=$\frac{TP}{TP+FN}$×100%
F1=2$\frac{RP}{R+P}$×100%
式中:TN代表正确识别为负类的负样本数;TP表示被准确划分为正类的正样本数;FN表示指正样本错误地被归类为负类的情况数;FP表示负样本被误判为正类的实例数量。
在对CNN-LSTM模型进行了300轮的训练过程中,追踪了模型损失值的变化,并通过分析训练与测试损失的曲线变化来评估模型的拟合质量,如图6所示。训练结果表明,模型的损失值降至2.1%,同时,模型的识别精度达到了92.37%,这进一步证实了CNN-LSTM模型展现出的优良收敛性。
图7所示为提出的4种分类模型识别方案的结果对比。根据图表数据,CNN-LSTM模型在各项性能指标上表现出了显著成绩。具体而言,模型的准确率达到了92.37%、精确率为91.10%、召回率为92.23%,以及F1为91.66%。相比之下,CNN-LSTM模型的准确率比CNN模型高2.16%,比LSTM模型提升了8.34%,且较SVM模型增加了11.16%。在精确率、召回率和F1 3个指标上,与CNN模型相比,CNN-LSTM模型分别提高了2.16%、5.27%、3.72%;与LSTM模相比,分别提升了4.46%、9.86%、7.21%;与SVM模型相比,增幅分别为8.72%、13.26%、11.02%。这些对比结果表明,CNN-LSTM模型在下肢动作识别方面比CNN、LSTM以及SVM模型具有更高的适应性和优越性。
为了比较不同的下肢动作识别模型的性能,将CNN、LSTM、CNN-LSTM、SVM等模型使用相同测试数据进行测试。在处理不同的下肢动作时,首先计算了测试样本被分类为特定类别的概率,随后基于这些概率,对各个阈值条件下的假正率与真正率进行了计算,并据此绘制了相应的ROC曲线。针对每种下肢动作,在各个测试集中分别生成了对应的ROC曲线。通过对每种动作识别模型产生的ROC曲线进行平均处理,获得了每种模型针对下肢动作识别任务的综合ROC曲线,如图8所示。
图9所示为通过4种不同模型的混淆矩阵,展示了对下肢动作进行四分类的结果,明显展现出所提出的CNN-LSTM模型能够从下肢7通道的表面肌电图信号中,准确地识别出各类下肢动作。在混淆矩阵中,对角线元素代表每个动作正确分类的样本数,而非对角线元素揭示了分类错误的样本数。动作分类包括:0表示行走,1表示上楼梯,2表示下楼梯,以及3表示蹲起动作。
(1)由于CNN模型对于较长序列的肌电信号的全局信息捕捉能力较差,结合LSTM,提出了CNN-LSTM模型。该模型能够捕捉全局信息和序列关系,解决了CNN模型局部特征信息缺失的问题。
(2)本研究构建并评估了4种不同的动作识别模型,旨在识别下肢康复训练中的4种动作。从20位参与者收集的sEMG中,提取了MAV、RMS、MPF和MDF4个关键特征值,应用于机器学习模型的输入。将预处理后的sEMG数据应用于CNN、LSTM、SVM模型的学习与训练,本研究比较了各模型在下肢动作识别任务上的效果。实验结果显示,CNN-LSTM模型在测试数据上的平均准确率达到了92.37%。
因此,本研究提出的CNN-LSTM模型相较于其他模型展现出显著的优势,验证了其在下肢动作识别领域的卓越性能。在未来工作中,本研究计划将参与者群体从健康人群扩展到下肢功能受损的患者,并进一步规范肌电信号的采集流程,以此提高基于sEMG的下肢康复机器人在不同患者群体中的动作识别准确率。此外,还将探索和改进不同的训练模型,以优化动作识别的性能。
  • 国家自然科学基金(52365039)
  • 自治区天山英才项目(2023TSYCLJ0051)
参考文献 引证文献
排序方式:
[1]
Yang C, Xi X G, Chen S J, et al. SEMG-based multifeatures and predictive model for knee-joint-angle estimation[J]. AIP Advances, 2019, 9(9): 095042.
[2]
Totty M S, Wade E. Muscle activation and inertial motion data for noninvasive classification of activities of daily living[J]. IEEE Transactions on Biomedical Engineering, 2018, 65: 1069-1076.
[3]
崔冰艳, 邓嘉, 张祥. 基于表面肌电信号的手势识别方法[J]. 科学技术与工程, 2023, 23(35): 15133-15141.
Cui Bingyan, Deng Jia, Zhang Xiang. Hand gesture recognition based on surface electromyography[J]. Science Technology and Engineering, 2023, 23(35): 15133-15141.
[4]
史小华, 卢浩, 廖梓宇, . 基于表面肌电信号的下肢康复主动训练[J]. 科学技术与工程, 2018, 18(17): 61-66.
Shi Xiaohua, Lu Hao, Liao Ziyu, et al. Active training of lower limb rehabilitation based on surface electromyography signal[J]. Science Technology and Engineering, 2018, 18(17): 61-66.
[5]
黄圣, 茅健. 基于改进深度森林的表面肌电手势识别方法[J]. 上海工程技术大学学报, 2023, 37(2): 190-197.
Huang Sheng, Mao Jian. Recognition method of sEMG gesture based on improved deep forest[J]. Journal of Shanghai University of Engineering Science, 2023, 37(2): 190-197.
[6]
Ai Q. Research on lower limb motion recognition based on fusion of sEMG and accelerometer signals[J]. Symmetry, 2017, 9(8): 147.
[7]
刘国平, 王南星, 周毅, . 基于改进ReliefF算法的哑铃动作识别[J]. 科学技术与工程, 2019, 19(32): 219-224.
Liu Guoping, Wang Nanxing, Zhou Yi, et al. Recognition of dumbbell movements based on an improved reliefF algorithm[J]. Science Technology and Engineering, 2019, 19(32): 219-224.
[8]
赵世昊, 周建华, 伏云发. 注意力机制CNN结合肌电特征矩阵的手势识别研究[J]. 电子测量与仪器学报, 2023, 37(6): 59-67.
Zhao Shihao, Zhou Jianhua, Fu Yunfa. Investigation of gesture recognition using attention mechanism CNN combined electromyography feature matrix[J]. Journal of Electronic Measurement and Instrumentation, 2023, 37(6): 59-67.
[9]
Xi X G, Yang C, Shi J H, et al. Surface electromyography-based daily activity recognition using wavelet coherence coefficient and support vector machine[J]. Neural Processing Letters, 2019, 50(3): 2265-2280.
[10]
刘光达, 许蓝予, 肖若兰, . 基于FGNN算法的sEMG肌肉疲劳分类方法[J]. 科学技术与工程, 2022, 22(19): 8370-8377.
Liu Guangda, Xu Lanyu, Xiao Ruolan, et al. Muscle fatigue classification method of sEMG signal based on FGNN algorithm[J]. Science Technology and Engineering, 2022, 22(19): 8370-8377.
[11]
Sui X W, Wan K X, Zhang Y. Pattern recognition of sEMG based on wavelet packet transform and improved SVM[J]. Optik, 2019, 176: 228-235.
[12]
刘声中, 许德章. 一种基于改进sEMG活动段检测的手部运动意图识别新方法[J]. 嘉兴学院学报, 2022, 34(6): 105-112.
Liu Shengzhong, Xu Dezhang. A new method for hand motion intent recognition based on improved sEMG signal active segment detection[J]. Journal of Jiaxing University, 2022, 34(6): 105-112.
[13]
Gupta R, Agarwal R. Electromyographic signal-driven continuous locomotion mode identification module design for lower limb prosthesis control[J]. Arabian Journal for Science and Engineering, 2018, 43(12): 7817-7835.
[14]
Reaz M, Hussain M, Mohd-Yasin F. Techniques of EMG signal analysis: detection, processing, classification and applications (Correction)[J]. Biological Procedures Online, 2006, 8(1): 11-35.
[15]
Xu H, Xiong A. Advances and disturbances in sEMG-based intentions and movements recognition: a review[J]. IEEE Sensors Journal, 2021, 21(12): 13019-13028.
[16]
Chiang J, Wang Z J, & McKeown M. A hidden Markov, multivariate autoregressive (HMM-mAR) network framework for analysis of surface EMG (sEMG) data[J]. IEEE Transactions on Signal Processing, 2008, 56(8): 4069-4081.
[17]
王睿, 杨建荣. 基于长短时记忆神经网络易损性分析的适用性研究[J]. 四川建筑科学研究, 2024, 50(2): 9-15.
Wang Rui, Yang Jianrong. Study on the applicability of vulnerability analysis based on long short-term memory neural network[J]. Sichuan Building Science, 2024, 50(2): 9-15.
2025年第25卷第7期
PDF下载
156
76
引用本文
BibTeX
文章信息
doi: 10.12404/j.issn.1671-1815.2403062
  • 接收时间:2024-04-25
  • 首发时间:2026-03-30
  • 出版时间:2025-03-08
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2024-04-25
  • 修回日期:2024-06-05
基金
国家自然科学基金(52365039)
自治区天山英才项目(2023TSYCLJ0051)
作者信息
    1 新疆大学智能制造现代产业学院, 乌鲁木齐 830017
    2 新疆医科大学第一附属医院, 医学检验中心, 乌鲁木齐 830054

通讯作者:

* 陶庆(1978—),男,汉族,江苏常熟人,博士,教授。研究方向:先进设计与制造、虚拟现实技术。E-mail:
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/kxjsygc/CN/10.12404/j.issn.1671-1815.2403062
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数

Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)

Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏