基于表面肌电信号的手指运动模式识别系统

Recognition system of finger movement pattern based on sEMG

目的:建立一套基于表面肌电信号(sEMG)的手指运动模式识别系统。方法:研发了基于有源电极的sEMG检测装置和包含信号采集、分析、离线识别和在线识别模块的配套软件。结果:该系统能够同时采集4个通道sEMG数据,用16种时域、频域和时频域方法对其分析,并可以对7种sEMG特征提取及分类方法进行对比研究,还能在线控制假手。结论:本系统为多指、多自由度假手sEMG控制方法的研究提供了一个实验平台,该系统不仅可以用于手指运动模式识别,而且也可以用于人体其他运动模式的sEMG特征提取方法和分类器的对比研究。

Objective： To construct a recognition system of finger movement pattern based on sEMG. Method： A sEMG-detection device based on active electrodes were developed and set up a support software system including signal-sampling, signal-analysis, off-line recognition and on-line recognition modules. Result： Four channels sEMG signals can be sampled simutaneously and be analyzed by 16 methods of time domain （TD）, frequency domain and time-frequency domain. The system can also make a contrastive research on 7 approaches of sEMG features extraction and classification. Moreover, it has the ability of controlling the multi-fingered robot hand on line. Conclusion： The system provides an experimental platform for the sEMG researches on multi-fingered and multi- DOF prosthetic hand controlling. It can not only be used to recognize the finger movement pattern but can also be used to extracting sEMG features from human＇s other body movement pattern and doing contrastive study on classifier.