华中科技大学学报（自然科学版）2013，Vol.41Issue(4)：75-79,5.
基于在线SVM的自适应sEMG人机交互系统
Adaptive myoelectric human-machine interface systems using online support vector machines
摘要
Abstract
To improve the adaptive ability of the man-machine interface (HMD system, an adaptive surface electromyographic signal (sEMG) based HMI system was proposed by online support vector machine (SVM). Control data samples were updated and trained by incremental online learning algorithm in real-time operation. Meanwhile, the image information of eyes was employed as the feedback information of the system. Adaptive boosting (Adaboost) algorithm was adopted to recognize whether left or right eye was closing, and the recognition result was used to decide which sEMG signal was updated as training sample of SVM. The presented system can adjust itself in real time operation. Experiments results show that the visual information makes the system effectively avoid misclassification because single sEMG signal lack of credible information. Furthermore, sEMG HMI system keeps a stable performance in its long-term operation.关键词
自适应sEMG(表面肌电信号)/人机交互/支持向量机/闭眼状态监测/Adaboost算法Key words
adaptive surface electromyographic sygnal (sEMG)/ HMI/ SVM/ eyes closing monitoring/ Adaboost algorithm分类
信息技术与安全科学引用本文复制引用
张毅,许新丽,罗元..基于在线SVM的自适应sEMG人机交互系统[J].华中科技大学学报（自然科学版）,2013,41(4):75-79,5.基金项目
科技部国际合作项目(2010DFA12160). （2010DFA12160）
