南京理工大学学报（自然科学版）2023，Vol.47Issue(6)：740-747,8.DOI:10.14177/j.cnki.32-1397n.2023.47.06.002
基于惯性信号与肌电信号的手势识别方法
Gesture recognition method based on inertial signal and electromyography
王胜东 1李忠新1
作者信息
- 1. 南京理工大学 机械工程学院,江苏 南京 210094
- 折叠
摘要
Abstract
In order to solve the problem of low accuracy of gesture recognition caused by the large error of human surface electromyography(EMG),a new gesture feature set is extracted by fusing the inertial motion signal of finger tips and joints relative to lumbar caudal vertebrae with the EMG signal of arm,and then the improved tabu search optimized back propagation(BP)neural network is used for classification and recognition.Experimental results show that compared with the traditional EMG feature set,the accuracy of the proposed feature set is improved by 8.5%,the accuracy of optimized BP neural network is 12.33%higher than that before optimization,and the comprehensive accuracy of gesture recognition can reach 99.75%.关键词
惯性信号/肌电信号/手势识别/运动信号/反向传播神经网络/改进禁忌搜索Key words
inertial signal/electromyography/gesture recognition/motion signal/back propagation neural network/improved tabu search分类
信息技术与安全科学引用本文复制引用
王胜东,李忠新..基于惯性信号与肌电信号的手势识别方法[J].南京理工大学学报（自然科学版）,2023,47(6):740-747,8.
