基于人工神经网络对sEMG信号的手势动作识别

针对利用表面肌电信号(sEMG)对手势动作的肌电信号的研究较少和sEMG信号处理过于复杂的问题,提出了利用人工神经网络和sEMG信号对人的手势动作进行识别研究,引入了MYO硬件设备对新的手势动作sEMG信号采集.利用MYO从手臂上获取每一个手势动作的sEMG信号,提取信号特征值,作为算法的训练数据和测试数据.采用人工神经网络中的反向传递神经网络算法来进行对4种不同手势动作分类,对应目标手指识别率在90.35%.研究结果可以被用来做临床诊断和生物医学的应用以及用于现代硬件的发展和更现代化的人机交互的发展.

Finger movements recognition based on artificial neural network on sEMG signal

Surface electromyography(sEMG)signals for fingers action research, which finger movements of the research methods is very less and signal processing is too complex.This experiment using the artificial neural network and sEMG signal to the person's finger gestures recognition research.The experiment introduced the MYO hardware equipment of the new finger movements sEMG signal acquisition.Use MYO from arm for each finger movements of sEMG signal,and then to extract the signal characteristic value as training data and test data of the algorithm, the recognition algorithm with artificial neural network of back propagation(BP)algorithm for the classification of four different finger gestures,its corresponding target finger recognition rate at 90.35%.The result can be used for clinical diagnosis and biomedical applications,the development of modern hardware and more modern the development of human-computer interaction.