基于肌电信号的前臂假肢动作识别研究与实现
发布时间：2018-08-19 15:25
【摘要】：前臂残疾者因为前臂和手部的缺失,身体功能和劳动能力受到极大影响。而前臂假肢通过工程方法为残疾者提供了人工假体,部分恢复了缺少的肢体功能。近年来,由于各项科学技术的发展,越来越多的商业精密前臂假肢被投放市场,而市场上的前臂假肢也变得越来越稳定,甚至出现了可以直接控制单根手指的假肢产品。这些假肢通过电极界面采集人体的表面肌肉电信号(surface electromyography,s EMG),处理后形成控制指令,实现残疾人通过残肢发出具体动作的肌肉电信号指令,控制高精密假肢完成具体动作。但目前市场上的精密假肢普遍价格较高,而实验室肌电处理方法主要测试于计算机仿真平台上,离实际使用仍有一定差距。为解决上述问题,本文尝试设计了一套基于开源嵌入式平台的智能假肢肌电信号采集与动作识别系统,可实现对人体前臂表面肌电信号的采集、处理并产生控制信号用于高精度智能假肢。本设计主要包括以下方面:1.根据人体肌肉运动单元和肌电信号产生原理,研究了目前肌电信号采集和预处理领域的相关方案。设计了表面肌电信号采集系统、电极位置和数字信号预处理方法。2.以模式识别技术为基本原理,介绍了表面肌电信号处理中特征提取的相关特征模型和求取方法,分析对比了一种综合时域和自回归特征的信号特征提取方法与另一种结合时域与功率谱描述的特征提取方法。同时介绍了使用的降维算法和分类器的数学原理及实现方法。通过实验分析了各方法的优劣。3.在已有方法基础上结合有限状态机和模式识别方法,提出了一套适用于嵌入式系统的肌电信号处理与模式识别新方法:FSM-TSD,该方法对大量的分类问题按不同状态进行了拆分,降低了分类难度,提高了分类准确率。4.基于商品化的肌电假肢控制相关要求,提出了一套肌电信号采集与手势识别系统的设计方案,达到对采集性能与市场参数的平衡。阐述了各种算法在资源有限的嵌入式微控制器平台上的实现方法。使用搭建的嵌入式平台进行肌电信号采集与动作识别功能的实验。
[Abstract]:The body and labor ability of the forearm disabled is greatly affected by the loss of the forearm and the hand. The forearm prosthesis provides artificial prosthesis for the disabled by engineering, partly restoring the missing limb function. In recent years, due to the development of various science and technology, more and more commercial precision forearm prostheses have been put on the market, and the forearm prostheses on the market have become more and more stable. These prostheses collect surface electromyographys (EMG),) signals of human body through electrode interface to form control instructions, and realize that disabled people can send out specific EMG signals through residual limbs, and control the high-precision prosthesis to complete the specific actions. However, the price of precision prosthesis in the market is generally high, and the method of electromyography in laboratory is mainly tested on the computer simulation platform, which is still far from the actual use. In order to solve the above problems, this paper attempts to design a set of intelligent EMG signal acquisition and motion recognition system based on open source embedded platform, which can realize the acquisition of EMG signal on the forearm surface of human body. Process and generate control signals for high-precision intelligent prostheses. This design mainly includes the following aspects: 1. 1. According to the principle of human muscle motility unit and EMG signal generation, the related schemes in the field of EMG signal acquisition and preprocessing are studied. The surface EMG signal acquisition system, electrode position and digital signal preprocessing method. 2. 2. Based on the principle of pattern recognition, the feature models and methods of feature extraction in surface electromyography (EMG) processing are introduced. A signal feature extraction method combining time domain and autoregressive features is analyzed and compared with another feature extraction method combining time domain and power spectrum description. At the same time, the mathematical principle and realization method of dimensionality reduction algorithm and classifier are introduced. The advantages and disadvantages of each method are analyzed by experiments. On the basis of existing methods, a new EMG signal processing and pattern recognition method named: FSM-TSDs is proposed, which combines finite state machine and pattern recognition method for embedded systems. This method splits a large number of classification problems according to different states. It reduces the difficulty of classification and improves the accuracy of classification. 4. Based on the commercial requirements of myoelectric prosthesis control, a design scheme of EMG signal acquisition and gesture recognition system is proposed to achieve the balance between acquisition performance and market parameters. The implementation of various algorithms on the platform of embedded microcontroller with limited resources is described. The embedded platform is used to carry out the experiment of EMG signal acquisition and motion recognition.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R496;TN911.7
本文编号：2192071
[Abstract]:The body and labor ability of the forearm disabled is greatly affected by the loss of the forearm and the hand. The forearm prosthesis provides artificial prosthesis for the disabled by engineering, partly restoring the missing limb function. In recent years, due to the development of various science and technology, more and more commercial precision forearm prostheses have been put on the market, and the forearm prostheses on the market have become more and more stable. These prostheses collect surface electromyographys (EMG),) signals of human body through electrode interface to form control instructions, and realize that disabled people can send out specific EMG signals through residual limbs, and control the high-precision prosthesis to complete the specific actions. However, the price of precision prosthesis in the market is generally high, and the method of electromyography in laboratory is mainly tested on the computer simulation platform, which is still far from the actual use. In order to solve the above problems, this paper attempts to design a set of intelligent EMG signal acquisition and motion recognition system based on open source embedded platform, which can realize the acquisition of EMG signal on the forearm surface of human body. Process and generate control signals for high-precision intelligent prostheses. This design mainly includes the following aspects: 1. 1. According to the principle of human muscle motility unit and EMG signal generation, the related schemes in the field of EMG signal acquisition and preprocessing are studied. The surface EMG signal acquisition system, electrode position and digital signal preprocessing method. 2. 2. Based on the principle of pattern recognition, the feature models and methods of feature extraction in surface electromyography (EMG) processing are introduced. A signal feature extraction method combining time domain and autoregressive features is analyzed and compared with another feature extraction method combining time domain and power spectrum description. At the same time, the mathematical principle and realization method of dimensionality reduction algorithm and classifier are introduced. The advantages and disadvantages of each method are analyzed by experiments. On the basis of existing methods, a new EMG signal processing and pattern recognition method named: FSM-TSDs is proposed, which combines finite state machine and pattern recognition method for embedded systems. This method splits a large number of classification problems according to different states. It reduces the difficulty of classification and improves the accuracy of classification. 4. Based on the commercial requirements of myoelectric prosthesis control, a design scheme of EMG signal acquisition and gesture recognition system is proposed to achieve the balance between acquisition performance and market parameters. The implementation of various algorithms on the platform of embedded microcontroller with limited resources is described. The embedded platform is used to carry out the experiment of EMG signal acquisition and motion recognition.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R496;TN911.7
【参考文献】
相关期刊论文 前7条
1 丁其川;熊安斌;赵新刚;韩建达;;基于表面肌电的运动意图识别方法研究及应用综述[J];自动化学报;2016年01期
2 彭亮;侯增广;王卫群;;康复机器人的同步主动交互控制与实现[J];自动化学报;2015年11期
3 张良清;周慧;杨琳;黄剑平;杨万章;李光林;;目标肌肉神经分布重建大鼠模型及低频电刺激的效果研究[J];中国康复医学杂志;2015年01期
4 王涛;侯文生;吴小鹰;万小萍;郑小林;;用于肌电假肢手控制的表面肌电双线性模型分析[J];仪器仪表学报;2014年08期
5 赵燕潮;;中国残联发布我国最新残疾人口数据[J];残疾人研究;2012年01期
6 游淼;邹国栋;林婉华;余龙;;基于肌动图与肌电图信号的假肢控制系统的研究[J];北京生物医学工程;2011年06期
7 贾晓枫,陈统一,陈中伟,张键,张晓文,斯扬,胡天培,高忠华,杨煜普;中国首例人体残肢神经信息控制电子假肢实验研究快报[J];中华物理医学与康复杂志;2004年01期
相关会议论文 前1条
1 罗永昭;孙为;;国内外现代实用假肢的进展和应用[A];第20届中国康协肢残康复学术年会论文选集[C];2011年
相关硕士学位论文 前1条
1 张志勇;肌电信号采集与肌电假肢控制的研究[D];哈尔滨工业大学;2010年
，本文编号：2192071
本文链接：https://www.wllwen.com/linchuangyixuelunwen/2192071.html
最近更新
教材专著
