为对手过头作业中的上肢肌肉疲劳状态进行有效识别,结合复杂装备的维修任务设计了手过头作业试验.通过采集被试的表面肌电信号(Surface Electromyography,SEMG)和主观疲劳状态及研究SEMG信号的时域、频域、非线性及参数模型特征计算方法,基于支持向量机(Support Vector Machine,SVM)采用核主成分分析(Kernel Principal Component Analysis,KPCA)进行特征降维并对手过头作业的肌肉疲劳状态进行识别.研究结果表明:手过头作业中斜方肌的SEMG贡献率最高;KPCA-SVM对训练集和测试集的疲劳识别率分别为0.998 27和0.832 18,与其他疲劳识别算法相比具有优越性.
Research on fatigue state recognition of upper limb muscles in overhead work
To effectively recognize the state of upper limb muscle fatigue during overhead work,an overhead work experiment was designed in conjunction with complex equipment maintenance tasks.Time-domain,frequency-domain,nonlinear,and paramet-ric modeling characteristics computational method of the SEMG signals were studied through the collection of surface electromyo-graphy(SEMG)signals and subjective fatigue levels.Kernel principal component analysis(KPCA)was employed to perform di-mension reduction and recognization of muscle fatigue status during overhead work.The findings revealed that the trapezius mus-cle exhibited the highest contribution rate in terms of SEMG during overhead tasks.The recognition rates for fatigue using KPCA-SVM were 0.998 27 and 0.832 18 for training set and test set,demonstrating superiority over other fatigue identification algorithms.
ergonomicsoverhead workmuscle fatiguesurface electromyographykernel principal component analysissupport vector machine
NETL
NSTL
万方数据
