摘要
肌电信号(Electromyography,EMG)是众多肌纤维中运动单元动作电位在时间和空间上的叠加,表面肌电信号(Surface Electromyography,sEMG)则是浅层肌肉EMG和神经干上电活动在皮肤表面的综合效应,能在一定程度上反映神经肌肉的活动。表面肌电信号作为一种无痛苦、无创伤且便捷的肌电检测方式,被广泛应用于手势识别、康复医疗和人机交互控制等领域,其中将肌电信号作为控制源的关键是根据肌电信号的差异能够识别出不同的手势动作。在MYO臂环采集人体上肢小臂处肌电信号后,从中提取平均绝对值、过零点数、波形长度3个时域特征,通过经粒子群优化算法(Particle Swarm Optimization,PSO)优化后的支持向量机(Support Vector Machine,SVM)分类器对采集到的sEMG进行分类识别,实验结果表明,优化后的SVM分类效果更好,平均识别率达到97.3%。
The Electromyography(EMG) signal is the superposition of the action potential of the motional unit in many muscle fibers in time and space,while the Surface Electromyography(SEMG) signal is the comprehensive effect of the EMG of the superficial muscle and the electrical activity of the nerve stem on the skin Surface,which can reflect the activity of the nerve muscle to a certain extent. As a painless,non-traumatic and convenient EMG detection method,surface EMG signal has been widely used in gesture recognition,rehabilitation medicine,human-computer interaction control and other fields. The key to use EMG signal as the control source is to judge different gestures according to the EMG signal. In the human upper limb MYO armlets acquisition forearm muscle after the electrical signals,average absolute value,zero crossings,waveform length are extracted,The collected SEMG recognition is classified by Particle Swarm Optimization algorithm,Particle Swarm Optimization(PSO) optimized Support Vector Machine(Support Vector Machine,SVM) classifier,the experimental results show that the optimized SVM classification effect is better,the average recognition rate is 97.3%.
引文
[1] Reaz M B I,Hussain M S,Mohd-Yasin F. Techniques of EMG signal analysis:detection,processing,classification and application[J].Biological Procedures Online,2006,8(1):11-35.
[2]傅桢茵,陆通,归丽佳,等.基于MYO的无线肌电控制鼠标[J].长春理工大学:自然科学版,2018,41(4):132-136.
[3]于擎,杨基海,陈香,等.基于BP神经网络的手势动作表面肌电信号的模式识别[J].生物医学工程研究,2009,28(1):6-10.
[4]马正华,乔玉涛,李雷,等.基于LDA的表面肌电信号分类研究[J].计算机工程与科学,2016,38(11):2321-2327.
[5]王杜,桂凯,曹红升,等.基于肌电控制的辅助型机器人外骨骼[J].燕山大学学报,2018,42(3):219-224.
[6]俞毕洪.基于表面肌电信号的穿戴式人体行为识别系统[D].浙江:浙江大学,2018.
[7] Chu J U,Moon I,Lee Y J, et al. A supervised feature-projection-based real-time EMG pattern recognition for multifunction myoelectric hand control[J].IEEE/ASME Transaction on mechatronics,2007,12(3):282-290.
[8] Cortes C,Vapnik V. Support-vector networks[J].Machine learning,1995,20(3):273-297.