下肢表面肌电信号的降维和映射分析

肌电信号的采集和分析是外骨骼式康复机器人关节预测控制的重要基础之一.肌电信号数据量大并且复杂,相关性较高,信号处理通用性和高效性低,分析和预测人体运动信息误差大.采用最大自主等长收缩标准化处理算法,大大提高了表面肌电信号的通用性和泛化能力,并基于主成分分析方法,对肌电信号降维处理,利用神经网络实现与下肢的映射分析.实验结果表明,通过对比分析不同的降维处理方式,主成分降维后处理的肌电信号平均相关性达0.93,利用神经网络预测人体正常行走的下肢三关节运动角度,具有良好的可重复性和较高的精度,可以实现人体下肢肌电信号和各关节的映射控制.     

二维相空间中肌电信息熵的计算

肌电信号的检测与分析对于临床诊断和康复工程具有重要的现实意义。从生理学角度上讲各种动作是肌肉的收缩产生的,同时伴有肌肉内部各运动单位电位的变化产生了肌电信号。通过对肌电信号一维时间序列的研究,发现其具备一定的混沌特性。信息熵是一个信息量的概念,在重构后的二维相空间中计算了各种肌电信号的信息熵,并对所得结果进行分析,确定了肌肉与动作之间的规律,此规律能为假肢的研究提供科学依据。     

基于眼肌肌电信号分析的眨眼识别

通过提取与分析人体的肌电信号,能有效地获取生物动作信息,帮助瘫痪病人恢复某些特定的生物功能.对上眼提肌肉的肌电信号进行分析与研究,提出了眨眼行为判断识别装置的设计过程.采用优化算法以及使用相关模拟软件(SIMULINK[2])、电路模型和选取恰当的神经元,达到有效地减弱外界噪声对肌电信号的干扰,省去小波建模分析,并在实验中取得了高准确度的判断效果.     

基于表面肌电的下肢肌肉功能网络构建及其应用研究

在肌电控制下肢康复辅具研究中,合适的肌电采集位置是运动模式识别的前提与基础.针对目前肌电采集位置缺乏成熟理论依据和统一标准的问题,选取90个下肢肌电采集点作为节点,通过计算节点间的肌电相关性,构建下肢肌肉功能网络,证明其具有小世界特性.实验结果表明:不同运动模式的网络具有明显的拓扑结构差异,通过网络特性分析可以确定与模式关联度大的肌电采集位置,取得较好的运动模式识别结果.通过构建及分析下肢肌肉功能网络,深入了解下肢运动模式更替过程中的肌肉协同工作机制,为下肢康复辅具控制中肌电采集位置的确定提供了理论支持.     

基于肌电信号的行走步态周期识别方法与实验系统的设计

通过下肢表面肌电信号的人体行走步态周期识别方法设计了实验系统;针对表面肌电信号的微弱性、交变性、低频性等特点,提出了识别肌肉动作起始时刻的峰—谷线性插值分段积分算法,并将该算法与阈值法相结合,提取足跟着地前肌肉动作起始时刻,从而达到划分步态周期的目的;该方法仅需单通道信号作为信息源,不同被测者可以选用不同的肌电信号;有效回避了肌电信号传感器零点漂移现象;文中分别对5位被测者行走时的7个下肢肌电信号进行采集,以VisualC++为工具,基于用户界面设计了步态周期的识别系统,其系统识别结果验证了该方法具有广泛性、可靠性、准确性和实用性。     

基于肌张力特征信息检测的痉挛传感器设计与应用

为脑卒中患者下肢肌肉痉挛检测及其评定,研制了一种基于肌张力特征信息进而评定痉挛等级的传感器,探讨了痉挛检测及量化评定标准需求,对痉挛传感器的结构进行了设计,针对不同痉挛等级的特征分布情况,基于痉挛特征信息,在改良Ashoworth量表MAS(Modified Ashworth Scale)基础上,进行了下肢关节被动活动过程中肌张力数据的采集实验。相关实验数据有效验证拟定的相关指标:包括肌张力分级阈值,痉挛等级等。实验结果表明痉挛传感器设计可行。     

人体肌电信号的生理层模型

研究表面肌电信号模型,能够更精确地描述肌肉在活动时所产生的肌电信号的形成过程.因此表面肌电信号模型的建立,可以实现对神经肌肉的控制,以及肌电信号的产生等问题的理解,为其特定参数值的提取,以及信号的实用性和真实性的分析打下了基础.在考虑了影响人体肌电信号主要参数值的特征条件后,通过matlab仿真软件建立一个能够仿真运动单位在不同激励下的表面肌电信号.目前的生理层仿真模型参数不可调,使得模型的精准性较差,而上述的改进和优化实现了参数的可调控性,使得模型更具实用性.     

人体上肢运动表面肌电特征研究

目的根据检测的人体上肢表面肌电信号,研究日常生活活动中男性、女性和左右上肢肌肉功能状态的差异性.方法 对20个被试者的三角肌中部、肱二头肌、肱三头肌、掌长肌和指伸肌五块肌肉进行肌电采集,每个动作重复3次,采用统计处理分析软件SPSS 14.0对时域特征参数平均振幅和积分肌电进行数据处理.结果女性的数值大于男性的数值;左上肢的数值大于右上肢的数值,经t检验知,左右上肢不存在无显著性差异.结论我们可以得到正常人体上肢日常生活活动中的肌电信号数值,为康复治疗、临床诊断、人体工效学提供评定指标.     

基于相干性的多频段脑肌电信号双向耦合分析

针对皮层肌肉相干性分析时不能确定耦合方向的局限性,根据神经肌肉信息的双向传递性,提出利用不同大脑功能区的脑电信号和动作相关的肌电信号,实现了相干函数对脑肌电信号的双向耦合分析.本文对不同握力模式下同步采集的脑肌电信号进行了多频段耦合分析.通过下行(EEG—>EMG)和上行(EMG—>EEG)分析发现,随着握力的增大,EEG能量、相干幅值和耦合强度均向高频段转移.与基于新型格兰杰因果关系的耦合方法进行比较,验证了相干性方法进行皮层肌肉双向耦合分析的可行性和优势.研究结果为探索基于皮层肌肉相干性的双向手部运动信息解码和上肢运动功能障碍分析提供了依据.     

利用多源运动信息的下肢假肢多模式多步态识别研究

运动状态识别对智能下肢假肢的控制非常关键,本文利用下肢表面肌电信号、腿部角度和足底压力信号在运动模式和步态分析中的优势和特点,对下肢假肢的多模式多步态识别进行研究.通过建立下肢运动信息系统,获取下肢多源运动信息.先提取下肢肌电信号的小波包能量作为特征,建立多个HMM对下肢假肢的运动模式进行识别;再根据大小腿和膝关节的角...     

Effect of fatigue on the stationarity of surface electromyography signals

The estimation of muscle fatigue using surface electromyography (SEMG) is of high relevance to evaluate ergonomic risk factors in the occupational settings. Signal stationarity plays an important role while selecting appropriate SEMG signal processing method for fatigue evaluation. The Fourier algorithm based signal processing methods (mean or median frequency of power spectrum) rely on the assumption that the signal under investigation is stationary. Stationarity of SEMG signals and its association with fatigue is rarely studied in the ergonomics literature. Therefore, this study was aimed at understanding the effect of fatigue on the stationarity of the SEMG data. Ten participants performed 40 min of fatiguing upper extremity exertions and SEMG data were recorded from the right upper trapezius muscle. The SEMG data recorded under static and dynamic conditions at the beginning and at the end of fatiguing exertions were used in the analysis. The stationarity analysis was performed for five window sizes of 128, 256, 512, 768 and 1024 ms using modified reverse arrangement test. The results showed that the muscle fatigue reduced the stationarity of the SEMG signal under static and dynamic conditions. The relationship between the muscle fatigue and the stationarity of the SEMG signal was found to be significant at the window size of 512 ms. A significantly higher fatigue related decrease in the stationarity was observed during dynamic exertions compared to the static exertions.Relevance to industryThe findings from the current study illustrate that the stationarity of SEMG signals could be used to quantify muscle fatigue under static and dynamic task conditions. These findings are useful to the ergonomic practitioners in conducting muscle fatigue estimation using SEMG.     

非线性PCA在表面肌电信号特征提取中的应用

针对表面肌电信号的特点,提出了一种应用非线性主分量分析(PCA)提取表面肌电信号特征的新方法.该方法在表面肌电信号滤波的基础上,采用非线性PCA方法完成数据压缩,将多路表面肌电信号转换为一维的特征数据主元,并以主元曲线的形式输出特征提取结果.本文采用基于自组织神经网络的非线性PCA对手臂尺侧腕伸肌和尺侧腕屈肌的两路表面肌电信号进行主元提取,试验结果表明,四种手部运动模式(握拳、展拳、腕外旋、腕内旋)对应的表面肌电信号利用该方法处理后,得到的主元曲线具有很好的类区分性,依据所得主元曲线的形状特征可以有效地进行手部动作类别的识别.     

Do mechanical tests of glove stiffness provide relevant information relative to their effects on the musculoskeletal system? A comparison with surface electromyography and psychophysical methods

The main purpose of the present study was to test the construct validity of two mechanical tests of glove stiffness using a surface electromyography (SEMG) methodology that would allow estimating the effect of glove stiffness on forearm muscle activation during a standardized grip contraction. The mechanical tests [free-deforming multidirectional test (FDMT) and Kawabata Evaluation System for Fabrics (KESF)] were applied on 27 gloves covering a wide range of stiffness. In 30 human subjects, a psychophysical assessment of these gloves was also carried on in addition to the SEMG test. The results showed that the sensitivity of the different tests to glove stiffness differences was slightly better for the FDMT (75% sensitivity) than for the psychophysical assessment (72%), while the SEMG test showed much lower sensitivity (13-31%, depending on the muscle). The SEMG test was highly correlated to the psychophysical assessment (0.88-0.95, depending on the muscle tested), and the FDMT (0.88-0.94) and KESF (0.77-0.86) mechanical tests, showing the construct validity of mechanical tests, particularly for the FDMT. It was concluded that mechanical tests provide relevant information relative to the effect of glove stiffness on the musculoskeletal system of the forearm.     

Electromyography signal analysis using wavelet transform and higher order statistics to determine muscle contraction

Abstract: Electromyography gives an electrical representation of neuromuscular activation associated with a contracting muscle. The electromyography signal acquires noise while travelling though different media. The wavelet transform is employed for removing noise from surface electromyography (SEMG) and higher order statistics are applied for analysing the signal. With the appropriate choice of wavelet, it is possible to remove interference noise (denoise) effectively in order to analyse the SEMG. Daubechies wavelets (db2, db4, db5, db6, db8), symmlet (sym4, sym5) and the orthogonal Meyer (dmey) wavelet can efficiently remove noise from the recorded SEMG signals. However, the most effective wavelet for SEMG denoising is chosen by calculating the root mean square difference and signal-to-noise ratio values. Results for both root mean square difference and signal-to-noise ratio show that wavelet db2 performs denoising best out of the wavelets. Furthermore, the higher order statistics method is applied for SEMG signal analysis because of its unique properties when applied to random time series, such as parameter estimation, testing of Gaussianity and linearity, deterministic and non-deterministic signal detection etc. Gaussianity and linearity tests as part of higher order statistics are conducted to understand changes in muscle contraction and to quantify the effectiveness of the noise removal process. According to the results, the SEMG signal becomes less Gaussian and more linear with increased force.     

A novel electromyographic signal simulator for muscle contraction studies

Mathematical simulation has been widely used in biomedical and biological sciences. In the case of the surface electromyographic (SEMG) activity, some models have been proposed aiming to study muscle contraction strategies that are used during different tasks and conditions. Most of SEMG simulators are based on energy modulation of a Gaussian noise. This work proposes a novel simulator in which the user-defined parameters are associated with the motor units (MUs) recruitment and their firing rate. Comparison between the mean spectrum of real SEMG signals collected in isometric contraction of the muscle biceps brachii and the mean spectrum obtained from simulated SEMG signals showed a good agreement, pointing the proposed simulator seems to be capable to generate consistent electromyographic signals in time and frequency domains and that can be used in many studies, in particular in the evaluation of automatic methods aimed to detect muscular contraction.     

Designing an accessible low-cost interactive multi-touch surface

This paper presents an accessible interactive surface specifically designed for cerebral palsy users. The work context is an educational centre for disabled users, although the system presented can be used in other contexts. The main contributions are as follows: the table leaves enough leg room to fit in wheelchairs and has a half-circle radius cut-out in the top which lets users in wheelchairs sit closer to the tabletop. Table height can be easily adjusted to children or adults. The tabletop screen angle can be easily adjusted to adapt to different body postures and allow for user comfort. The tabletop does not compromise user safety as the surface is protected with strong Plexiglas. Furthermore, an assessment of accessibility has been carried out to analyse its appropriateness to fit users with different characteristics.     

下肢外骨骼研究进展及关节运动学解算综述

随着传感融合、移动计算、智能驱动等技术的发展以及研究者对人体运动中下肢重要生物力学功能认知的逐步深化,下肢外骨骼机器人作为一种与下肢并联,能为穿戴者行走助力的可穿戴智能设备愈发受到世界各研究机构的重视。本文根据下肢外骨骼的用途和结构详细综述了近年下肢外骨骼的研究进展,并借此对下肢外骨骼的未来发展进行展望。并针对下肢外骨骼在实时运动学检测与控制上对小型传感器的迫切需求,提出一种能够用于控制下肢外骨骼的基于惯性测量单元的人体下肢关节运动学测量与解算技术,在基于惯性测量的单自由度关节角度结算上得到较好结果。     

Multi-body sensor data fusion to evaluate the hippotherapy for motor ability improvement in children with cerebral palsy

Hippotherapy is a new rehabilitation therapy for children with cerebral palsy (CP). Although it has been proved to be effective in clinical research, a quantitative evaluation of such results is still lacking in previous studies. In this research, one method for evaluating the effectiveness of hippotherapy based on body sensor network (BSN) is proposed. The method adopts distributed magnetic, angular rate, and gravity (MARG) sensors to evaluate the gross motor function of CP children by multi-sensor data fusion algorithm, the comparison results with the golden standard optical capture system demonstrate the robustness of sensor fusion algorithm. Moreover, via tracking one-year’s hippotherapy projects, a pilot study was conducted for measuring and evaluating the motor coordination function and gait parameters of CP children. Finally, the rehabilitation of gross motor function for the chosen subjects under different treatment periods was evaluated through kinematic analysis. The results of our method show that some children’s symptoms, such as limb stiffness, poor joint range of motion, scissors gait, knee flexion gait, have been relieved. Thus, it provides an empirical basis for hippotherapy in the rehabilitation of motor function corresponding to CP children.     

Electromyographic and discomfort analysis of confined-space shipyard welding processes

This study examined muscle fatigue and discomfort in a confined-space welding operation at a shipyard. Surface electromyography (SEMG) was recorded from seven upper extremity and torso muscles of welders welding in a mock-up of the work environment. Following spectral transform of the SEMG data the percentage of the total signal power in the 10-30 Hz frequency band was compared over time during welding. For the conventional stick electrode welding process (SMAW) several muscles exhibited an increase in the percent of the total signal power in the low-frequency band. Fewer muscles exhibited this fatigue-related spectral density shift with a wire welding process (FCAW) the shipyard has considered adopting. This finding suggests that localized muscle fatigue may be reduced by a change to the wire welding process. Subjectively reported discomfort was generally low for both processes, but confirmed the finding that discomfort in the low back and shoulder regions is experienced in this welding operation.     

Integration of active pauses and pattern of muscular activity during computer work

Abstract

Submaximal isometric muscle contractions have been reported to increase variability of muscle activation during computer work; however, other types of active contractions may be more beneficial. Our objective was to determine which type of active pause vs. rest is more efficient in changing muscle activity pattern during a computer task. Asymptomatic regular computer users performed a standardised 20-min computer task four times, integrating a different type of pause: sub-maximal isometric contraction, dynamic contraction, postural exercise and rest. Surface electromyographic (SEMG) activity was recorded bilaterally from five neck/shoulder muscles. Root-mean-square decreased with isometric pauses in the cervical paraspinals, upper trapezius and middle trapezius, whereas it increased with rest. Variability in the pattern of muscular activity was not affected by any type of pause. Overall, no detrimental effects on the level of SEMG during active pauses were found suggesting that they could be implemented without a cost on activation level or variability.

Practitioner Summary: We aimed to determine which type of active pause vs. rest is best in changing muscle activity pattern during a computer task. Asymptomatic computer users performed a standardised computer task integrating different types of pauses. Muscle activation decreased with isometric pauses in neck/shoulder muscles, suggesting their implementation during computer work.