基于MYO臂环的假肢手控制技术研究

发布时间：2018-06-15 00:25

本文选题：表面肌电信号 + 模式识别　；参考：《上海师范大学》2017年硕士论文

【摘要】：表面肌电信号(surface electromyography,sEMG)是人体肌肉收缩时产生的生物电信号。随着国内外学者的不懈努力,sEMG已经被广泛应用于临床检测、康复工程以及假肢手控制等领域中。目前,基于sEMG的假肢手控制技术已然成为研究的热点。与传统传感器相比,MYO臂环具有不受场地限制、交互自然、穿戴方便以及性价比高等优点,非常适合用来控制假肢手。所以,本文的目的在于研究一种基于MYO臂环的肌电假肢手控制技术,通过算法实现对人手动作模式识别和人手抓取力的预测,并结合在PC端开发的假肢手肌电控制系统进行验证。本文主要研究工作如下:(1)人手动作模式识别研究。本实验采用六阶巴特沃斯带通滤波器对MYO臂环采集的sEMG进行预处理,并提取5种时域特征,采用PCA和BP神经网络相结合的方法对人手动作模式进行分类。实验结果表明,运用PCA将特征样本映射到20维时,人手动作模式的识别率可达99%。(2)人手抓取力预测技术研究。本实验选取绝对平均值(MAV)和均方根(RMS)作为特征,以抓取力的八个档次为输出,建立了基于BP神经网络的抓取力预测模型。实验结果表明,确定抓取力按照大小分档的平均识别率达到了93.83%,能够满足假肢手控制的基本要求。(3)假肢手肌电控制系统设计。设计了一套基于MFC的肌电控制系统,该系统能够采集并分析sEMG,进而获取人手动作模式的活动意图,且对手指抓取力进行实时预测,经串口对假肢手进行驱动控制。最终,应用该系统验证了本课题方案的可行性。该肌电控制系统根据sEMG实现人手动作模式的实时分类和抓取力的实时预测。所提取的手部动作意图和抓取力可转换成不同的控制命令,能够提供一种有效的基于生物电信号的人机交互模式。该系统主要创新性的工作在于将高性价比MYO臂环应用到假肢手的控制中,实现了人手动作模式和抓取力的在线控制,在线平均识别率可达92%。而且,该系统安装和使用方便、抗干扰能力强以及具有很高的可控性,可以很好地满足残疾人对假肢手控制的需求。
[Abstract]:Surface electromyography (EMG) is a bioelectric signal produced when human muscles contract. With the unremitting efforts of scholars at home and abroad, SEMG has been widely used in clinical detection, rehabilitation engineering and prosthetic hand control. At present, the control technology of prosthetic hand based on SEMG has become a hot spot. Compared with traditional sensors, MYO arm ring has the advantages of no limitation of site, natural interaction, easy to wear and high cost performance, so it is very suitable for the control of prosthetic hand. Therefore, the purpose of this paper is to study a myoelectric prosthetic hand control technology based on MYO arm ring. The EMG control system of prosthetic hand was developed on PC. The main work of this paper is as follows: 1. In this experiment, the sixth order Butterworth band-pass filter is used to preprocess the SEMG collected by MYO arm ring, and five time domain features are extracted, and the manual action pattern is classified by PCA and BP neural network. The experimental results show that when PCA is used to map feature samples to 20 dimensions, the recognition rate of manual action pattern can reach 99%. In this experiment, the absolute mean value (MAV) and RMS (mean square root) are selected as the characteristics, and the prediction model of grab force based on BP neural network is established with eight grades of grab force as the output. The experimental results show that the average recognition rate of grasping force is 93.833.It can meet the basic requirements of prosthetic hand control. A set of electromyoelectric control system based on MFC is designed. The system can collect and analyze sEMG, and then obtain the active intention of the manual movement mode, and predict the finger grasping force in real time, and carry on the drive control to the prosthetic hand through the serial port. Finally, the feasibility of the project is verified by using the system. According to SEMG, the EMG realizes the real-time classification of manual action mode and the real-time prediction of grasping force. The extracted hand motion intention and grip force can be converted into different control commands, which can provide an effective human-computer interaction mode based on bioelectric signals. The main innovative work of this system is to apply the MYO arm ring with high performance to price ratio in the control of prosthetic hand. The on-line control of manual movement mode and grip force is realized, and the on-line average recognition rate can reach 92%. Moreover, the system is easy to install and use, has strong anti-interference ability and has a high controllability, which can meet the needs of the disabled in the control of prosthetic hands.
【学位授予单位】：上海师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R496;TP273

【参考文献】