仿生机械假手的肌电控制及其力触觉感知反馈方法研究
发布时间：2019-01-19 10:48
【摘要】：假手是一类典型的人机交互设备,对于辅助手臂截肢患者恢复手部功能有着重要的作用。早期的假手以装饰性为目的,用以在外观上装饰残疾人缺失的肢体。现代智能假手以电动假手为主,它是生物医学、计算机科学、电子学、控制科学、机器人学等众多学科融合的产物,其中基于肌电信号控制的假手最受关注。本文在国家自然科学基金项目、江苏省科技计划项目的资助下,从提高肌电假手的可控性和可感知性出发,进行了仿生机械假手的肌电控制及其力触觉感知反馈方法研究。文章首先从机构设计、控制策略、信息感知等方面对国内外假手的研究现状进行了分析。在此基础上针对实际需求提出了假手的设计指标,设计了包括手爪开合机构、手腕旋转机构的两自由度机械假手并对设计的假手机构进行了详细分析。针对假手信息感知的需求,设计了机构传感器一体化的假手手指,设计的手指能够检测假手的握力以及被抓物体与手指的接触位置。在假手机构设计的基础上设计了假手的测量控制系统,包括硬件测控电路和上位机测控软件。针对手部运动意图的识别,文章在设计表面肌电信号传感器的基础上对肌电信号进行了采集和特征提取,采用支持向量机(SVM)进行了手部/腕部8个动作的识别分类、采用广义回归神经网络(GRNN)实现了手部握力、手部三维推拉力的估计。由于肌电信号存在显著的个体差异,实际使用时需要根据不同佩戴者的肌电信号对假手的控制参数进行调整,而且佩戴者往往需要经过一段较长时间的适应训练才能控制假手运动。针对这一问题,本文进行了肌电信号解耦自适应学习的动作识别方法研究。首先建立了双通道肌电信号的解耦模型用以消除双路信号间的重叠干扰,随后设计了肌电信号自适应学习器用以自适应调节两路肌电信号的归一化比例因子,从而消除肌电信号的个体差异,实现肌电信号的自适应动作识别。在此基础上设计了模糊神经网络PID控制器用于实现假手的比例控制。针对假手需要抓取不同物体且抓取过程中物体的特性参数可能发生变化的问题,本文进行了基于抓取对象刚度模糊观测的假手反演控制方法研究,采用李雅普诺夫(Lyapunov)理论设计反演控制器,实现自由空间内的速度控制和约束空间内的握力控制;为了均衡不同物体对控制器性能的影响,设计模糊观测器实时估计被抓握物体的刚度,并将其用于调整反演控制器的参数,实现了假手对不同物体的平稳抓取控制。信息感知能力是智能假手的一个重要特征,目前的假手多处在无感知状态或自感知状态,即假手没有信息感知能力或感知到的信息只服务于假手控制器本身而假手的佩戴者却无法感知到假手的状态。针对假手对于佩戴者缺乏感知反馈的问题,文章进行了具有力触觉感知反馈功能的假手控制方法研究。设计振动袖带用于向佩戴者反馈假手的状态信息,并基于振动触觉再现技术设计了三种力触觉信息的触觉编码,包括握力、手腕三维力、物体在假手上的滑动信息。
[Abstract]:The artificial hand is a typical human-computer interaction device, which plays an important role in the rehabilitation of the hand function of the amputee. The early artificial hand is used as a decorative purpose to decorate the missing limb of the person with disabilities in appearance. The modern intelligent artificial hand is mainly the electric artificial hand, it is the product of many subjects such as biomedicine, computer science, electronics, control science, robotics and so on, in which the artificial hand based on the myoelectric signal control is the most concerned. In this paper, under the support of the National Natural Science Foundation of China (NSFC) and the project of Jiangsu Province's Science and Technology Program, the author has made a study on the control of the mechanical prosthetic hand and the tactile sensation feedback of the bionic mechanical prosthetic hand from the viewpoint of improving the controllability and the perceptibility of the muscle electric artificial hand. In this paper, the present situation of the research on the home and abroad is analyzed from the aspects of the mechanism design, the control strategy, the information perception and so on. On the basis of this, the author puts forward the design index of the false hand for the actual demand, and designs the two-degree-of-freedom mechanical prosthetic hand including the hand-jaw opening-closing mechanism and the wrist-rotating mechanism and analyzes the designed prosthetic hand mechanism. Aiming at the demand of false hand information perception, a false hand finger integrated with the mechanism sensor is designed, and the designed finger can detect the holding force of the artificial hand and the contact position of the object to be grasped and the finger. The control system of the false hand is designed on the basis of the design of the artificial hand mechanism, including the hardware measurement and control circuit and the measurement and control software of the upper computer. In view of the recognition of the hand movement intention, on the basis of designing the surface myoelectric signal sensor, the electromyographic signal is collected and extracted, and the recognition and classification of the hand/ wrist 8 actions are carried out by using a support vector machine (SVM). In this paper, a generalized regression neural network (GRNN) is used to estimate the hand-holding force and the three-dimensional push-pull force of the hand. Due to the significant individual difference of the myoelectric signal, the control parameters of the artificial hand need to be adjusted according to the myoelectric signals of different wearers, and the wearer often needs to be trained for a long time to control the false hand movement. In order to solve this problem, this paper studies the motion recognition method of the decoupling self-adaptive learning of the myoelectric signal. firstly, a decoupling model of the two-channel muscle electric signal is established to eliminate the overlapping interference between the two-way signals, and the self-adaptive motion recognition of the myoelectric signal is realized. On this basis, the fuzzy neural network PID controller is designed to realize the proportional control of the false hand. In order to solve the problem that the artificial hand needs to grasp different objects and the characteristic parameters of the objects in the process can change, this paper studies the artificial hand inversion control method based on the object stiffness fuzzy observation, and adopts Lyapunov theory to design the inversion controller. in order to balance the influence of different objects on the performance of the controller, a fuzzy observer is designed to estimate the rigidity of the gripped object in real time and to adjust the parameters of the inversion controller, and the stable grabbing control of the artificial hand on different objects is realized. The information sensing capability is an important feature of the intelligent artificial hand, and the current false hand is in the non-sensing state or the self-sensing state, that is, the artificial hand has no information sensing capability or the perceived information is only serving the false hand controller, and the wearer of the artificial hand can not perceive the state of the false hand. In view of the problem of the false hand's lack of sense feedback to the wearer, the paper makes a study on the method of artificial hand control with the function of haptic perception feedback. The design vibration cuff is used to feedback the state information of the artificial hand to the wearer, and the haptic coding of three kinds of force tactile information is designed based on the vibration haptic reproduction technique, including the grip force, the three-dimensional force of the wrist, and the sliding information of the object on the artificial hand.
【学位授予单位】：东南大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP241
本文编号：2411298
[Abstract]:The artificial hand is a typical human-computer interaction device, which plays an important role in the rehabilitation of the hand function of the amputee. The early artificial hand is used as a decorative purpose to decorate the missing limb of the person with disabilities in appearance. The modern intelligent artificial hand is mainly the electric artificial hand, it is the product of many subjects such as biomedicine, computer science, electronics, control science, robotics and so on, in which the artificial hand based on the myoelectric signal control is the most concerned. In this paper, under the support of the National Natural Science Foundation of China (NSFC) and the project of Jiangsu Province's Science and Technology Program, the author has made a study on the control of the mechanical prosthetic hand and the tactile sensation feedback of the bionic mechanical prosthetic hand from the viewpoint of improving the controllability and the perceptibility of the muscle electric artificial hand. In this paper, the present situation of the research on the home and abroad is analyzed from the aspects of the mechanism design, the control strategy, the information perception and so on. On the basis of this, the author puts forward the design index of the false hand for the actual demand, and designs the two-degree-of-freedom mechanical prosthetic hand including the hand-jaw opening-closing mechanism and the wrist-rotating mechanism and analyzes the designed prosthetic hand mechanism. Aiming at the demand of false hand information perception, a false hand finger integrated with the mechanism sensor is designed, and the designed finger can detect the holding force of the artificial hand and the contact position of the object to be grasped and the finger. The control system of the false hand is designed on the basis of the design of the artificial hand mechanism, including the hardware measurement and control circuit and the measurement and control software of the upper computer. In view of the recognition of the hand movement intention, on the basis of designing the surface myoelectric signal sensor, the electromyographic signal is collected and extracted, and the recognition and classification of the hand/ wrist 8 actions are carried out by using a support vector machine (SVM). In this paper, a generalized regression neural network (GRNN) is used to estimate the hand-holding force and the three-dimensional push-pull force of the hand. Due to the significant individual difference of the myoelectric signal, the control parameters of the artificial hand need to be adjusted according to the myoelectric signals of different wearers, and the wearer often needs to be trained for a long time to control the false hand movement. In order to solve this problem, this paper studies the motion recognition method of the decoupling self-adaptive learning of the myoelectric signal. firstly, a decoupling model of the two-channel muscle electric signal is established to eliminate the overlapping interference between the two-way signals, and the self-adaptive motion recognition of the myoelectric signal is realized. On this basis, the fuzzy neural network PID controller is designed to realize the proportional control of the false hand. In order to solve the problem that the artificial hand needs to grasp different objects and the characteristic parameters of the objects in the process can change, this paper studies the artificial hand inversion control method based on the object stiffness fuzzy observation, and adopts Lyapunov theory to design the inversion controller. in order to balance the influence of different objects on the performance of the controller, a fuzzy observer is designed to estimate the rigidity of the gripped object in real time and to adjust the parameters of the inversion controller, and the stable grabbing control of the artificial hand on different objects is realized. The information sensing capability is an important feature of the intelligent artificial hand, and the current false hand is in the non-sensing state or the self-sensing state, that is, the artificial hand has no information sensing capability or the perceived information is only serving the false hand controller, and the wearer of the artificial hand can not perceive the state of the false hand. In view of the problem of the false hand's lack of sense feedback to the wearer, the paper makes a study on the method of artificial hand control with the function of haptic perception feedback. The design vibration cuff is used to feedback the state information of the artificial hand to the wearer, and the haptic coding of three kinds of force tactile information is designed based on the vibration haptic reproduction technique, including the grip force, the three-dimensional force of the wrist, and the sliding information of the object on the artificial hand.
【学位授予单位】：东南大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP241
【相似文献】
相关期刊论文 前5条
1 刘佳;宋爱国;;人手柔性触觉感知特性[J];东南大学学报(自然科学版);2007年05期
2 岳宏;基于虚拟现实触觉感知接口技术的研究与进展[J];机器人;2003年05期
3 汪莉;万华根;彭群生;;基于CHARMM力场的蛋白质分子场计算及触觉感知[J];计算机辅助设计与图形学学报;2009年07期
4 王辉静;;基于MEMS技术的三维触觉感知阵列研究[J];深圳信息职业技术学院学报;2008年02期
5 马梅若;;大数据预言[J];中国经济和信息化;2013年08期
相关会议论文 前1条
1 王宏伟;;虚拟触觉感知系统中构造变形体方法的研究[A];全国先进制造技术高层论坛暨制造业自动化、信息化技术研讨会论文集[C];2005年
相关重要报纸文章 前1条
1 ;反间计[N];中国包装报;2006年
相关博士学位论文 前2条
1 陈思;皮肤摩擦触觉感知的机理研究[D];中国矿业大学;2016年
2 吴常铖;仿生机械假手的肌电控制及其力触觉感知反馈方法研究[D];东南大学;2016年
相关硕士学位论文 前2条
1 吴雅丽;网络环境下模特呈现对消费者触觉感知的影响研究[D];湖南大学;2014年
2 张海涛;气动触觉感知再现装置的研究[D];哈尔滨工业大学;2007年
，本文编号：2411298
本文链接：https://www.wllwen.com/shoufeilunwen/xxkjbs/2411298.html
