基于步态检测算法的辅助行走系统设计

为了保证盲人和视障患者安全方便出行和减少误报警,设计一种基于步态检测算法的辅助行走系统。信号处理算法上利用小波包对信号进行降噪,提取三轴加速度的均值、标准差、方差与小波能量谱,并结合足底压力信息组成多维参数作为步态特征,选取基于遗传算法优化支持向量机进行步态识别。硬件部分主要包括三轴加速度信号和足底压力信号采集硬件设计。软件部分主要包括对足底压力信号和三轴加速度信号的采集和处理以及一款能够自动拨号安卓手机APP。实验结果表明,该系统基于多传感器融合的步态检测技术,步态检测识别率平均达90.48%。该系统便携性好、功耗低、测障效果好,在辅助行走领域具有一定的研究意义和实用价值。     

基于多源信息和广义回归神经网络的下肢运动模式识别

针对目前下肢运动模式识别率低的问题,提出了一种基于多源信息和广义回归神经网络(GRNN)的下肢运动模式识别方法.通过足底压力信息将人体日常下肢动作分解为不同的动作片段以组成识别目标集.采用下肢表面肌电信号中的偏度、峭度、功率谱熵,以髋关节角度作为腿部特征值,利用主成分分析(PCA)方法对文中提取的特征值进行降维处理,以缩短模型训练时间,防止过拟合.最后,利用GRNN对目标集中平地行走、上楼、下楼3种动作进行识别.实验结果表明,该方法的正确识别率为90.16%.     

基于隐马尔可夫模型的动力型下肢假肢步态预识别

为使动力型假肢膝关节协调配合人体的运动,关键是对人体行走步态进行有效预识别.本文利用安装在假肢接受腔上的加速度传感器和安装在足底的压力传感器采集人体的运动信息,根据人体运动的规律性和重复性特点,通过将隐马尔可夫模型引入到所获得的运动信息中来分析并预识别人体的运动步态.实验表明,基于隐马尔可夫模型的动力型下肢假肢的步态预识别方法是有效并且准确的.     

采用核主成分分析和相关向量机的人体运动意图识别

针对人体步态识别率低的问题,提出了一种将核主成分分析(KPCA)和相关向量机(RVM)相结合的步态识别方法.首先,选择表面肌电信号(s EMG)作为步态识别信息源,提取表面肌电信号的小波包能量特征.然后,采用核主成分分析方法降维特征值去除冗余的信息,得到能反映步态特征的特征值.最后,利用相关向量机对步态特征向量进行分类,识别平地行走、上楼、下楼、上坡、下坡5种步态.通过分析不同受试者步态识别结果,验证了该方法的可行性和实用性,并和BP(反向传播)神经网络、SVM(支持向量机)等方法比较,结果表明该方法分类时间为2.6609×10~(-4)s,识别正确率为96.67%,是一种有效的步态识别方法.     

仿人足底肌电特征的机器人行走规划

模仿人类行走规律是规划双足机器人运动的基础.以往模仿人类步态主要通过视觉方法或惯性模块测量(Inertia measurement unit, IMU)方法捕捉人体特征点轨迹.这些方法不考虑零力矩点(Zero moment point, ZMP)的相似性.为解决该问题,本文提出了一种基于足底肌电信号(Electromyography, EMG)和惯性模块测量信号的混合运动规划方法.该方法通过测量足底肌电信号计算出足底压力中心的位置以及踝关节扭矩,结合惯性模块所测量的人体躯干和双足轨迹,来规划双足机器人的步态.首先,用肌电仪测量足底肌电信号,用惯性测量模块测量人体各肢体部分的姿态轨迹,经数据标定后作为仿人机器人的运动参考; 然后,通过预观控制输出稳定的步态.为确保仿人行走的效果,基于人体相似性对运动数据进行了步态优化.实验验证和分析表明, EMG信号超前ZMP约160ms,利用这个特性实现了对压力点位置的有效预测,提高了机器人在线模仿人类行走的稳定性.     

基于实时意图识别的外骨骼机器人步态切换

根据外骨骼穿戴者的实时运动意图,切换外骨骼机器人的行走步态,是外骨骼机器人研究热点问题之一.针对外骨骼机器人能够根据穿戴者的运动意图控制步态切换问题,提出采集人体表面肌电信号进行解码,通过BP神经网络识别穿戴者的运动意图,实现外骨骼进行步态规划,利用关节角度与足底压力的相关性以及重心轨迹,对所规划的步态的稳定性进行分析,依据不同的运动意图完成步态进行实时切换,验证了基于实时意图识别控制的外骨骼步态切换的可行性.     

RVM算法在下肢辅助运动设备中的应用

相关向量机(RVM)是一种适用于非线性、较高维数的小样本事件的稀疏贝叶斯学习算法。笔者提出基于相关向量机利用足底压力信号预测人类行走时步态轨迹算法,通过相关压力传感器和磁栅尺同步采集足底部位的压力数据和踝关节角度数据,并利用Savitzky-Golay滤波算法滤波和主成分分析算法,目的在于减少压力信号维数,获得压力主分量信号。利用压力主分量信号进行最大期望算法(EM)训练,得出RVM模型。     

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...     

基于足底压力分布时空HOG特征的步态识别方法

提出一种基于足底压力分布时空HOG的步态识别算法,在特征层对足底压力的时间域和空间域信息进行融合。首先寻找足底总压力时间曲线上的极大值和极小值等几个特征点,利用这几个特征点所对应时刻的足底压力分布来构建时空HOG特征向量,最后采用SVM进行步态识别。采集不同行走速度下30人的单步足底压力分布数据进行实验,在不区分样本速度的情况下,该方法的识别率为93。5%。实验结果表明足底压力分布时空HOG特征能较好地刻画步态动力学特征,且具有良好的速度适应性。     

结合无符号Laplace谱特征的触觉步态识别算法

针对单纯利用压力点分布特征进行触觉步态识别的不足,提出了一种结合无符号Laplace谱特征的动态触觉步态识别算法。利用足底压力数字化场地采集常速、快速和慢速三种情况下的触觉步态数据,生成足底压力分布图像,并根据足底解剖学的结构划分区域;以足底压力图像各区域为节点构造结构图,并采用无符号Laplace矩阵表示;通过对该矩阵进行奇异值分解（Singular Value Decomposition,SVD）获取谱特征,并结合形状特征得到触觉步态特征;选择“一对一”的支持向量机（Support Vector Machine,SVM）多分类方法,按照人在行走过程中不同的速度分别构造分类器,从而实现动态触觉步态的识别。实验结果表明该识别算法对不同速度样本数据的触觉步态识别正确率都较高。     

基于几何与物理特征融合的智能下肢假肢运动意图识别

传统的意图识别方法所用传感器数量及种类较多,特征向量维数偏高,统计特征对短时样本具有不稳定性.将关节角表示的几何特征与加速度、角速度表示的物理特征有机融合并应用于智能下肢假肢的运动意图识别.首先,利用惯性测量单元于健侧大腿、小腿处采集的摆动相前期短时时序数据解算膝关节角,以获取大腿、小腿绕关节轴的转动特性;其次,对物理...     

基于改进模板匹配的智能下肢假肢运动意图实时识别

传统下肢假肢运动意图识别常使用多模态传感器采集残肢侧时频域特征,在短时意图识别中,具有一定的复杂性和滞后性,且时频域特征不稳定难以达到实时意图识别的目的.鉴于此,提出基于改进模板匹配技术的智能下肢假肢运动意图实时识别的方法.在重新定义单侧下肢截肢者的运动模式后,仅采用惯性传感器采集健肢侧位于摆动相的数据,基于改进的模板匹配,通过滑动窗口创建完备的模板库,使得每类运动模式在库中有充足的原子模式,对下肢假肢的运动意图进行实时识别.实验结果表明,所提出方法在5种稳态模式(平地行走、上下楼、上下坡)的识别率为99.50%,在引入8种转换模式后的识别率为97.03%,可以大大提高下肢假肢实时识别性能,助力单侧下肢截肢者更自然地行走.     

一种基于特征增强和决策融合的人体行为识别方法

利用三轴加速度传感器进行人体行为识别一直是传感器数据处理、模式识别领域的研究热点。加速度数据往往存在着多种动作数据难以区分的情况,特别是走、上楼、下楼这3个动作数据非常相似,这给正确识别这3种人体动作带来了较大的难度。提出一种基于特征增强与决策融合的行为识别方法,通过对部分特征值进行增强处理和对多个分类结果进行决策融合来识别走、上楼、下楼这些难以区分的相似动作。实验验证,所提方法可克服由于加速度数据的相似性而导致的动作识别正确率低、识别误差大的情况,有效提高人体行为识别率,且可在实际应用中实时识别人体行为动作。     

A comparison of foot/ground interaction during stair negotiation and level walking in young and older women

Stair design and environmental conditions may play a role in slip accidents on stairs in the workplace, but little is known about the slip resistance requirements on stairs compared to level walking. Older adults have an increased risk of falling compared to younger adults and may be at greater risk during stair negotiation. The purpose of this study was to determine whether the ground reaction force profiles and peak required coefficient of friction (RCOF) differed between young and elderly women or between stair ascent, stair descent and overground walking. While there was a trend towards less risky stair descent behaviour in the older women in terms of their peak RCOF values during stair descent, the increased vertical loading rate in the older women may imply reduced dynamic balance control. The largest mean RCOF peaks occurred during stair ascent in both young and older women, but there were several overground walking trials in both groups and a few stair descent trials of the young women, which resulted in RCOF peaks greater than 0.5. These results should be considered when choosing stair surface materials, particularly in occupational and outdoor settings where the tread surfaces may become wet or contaminated.     

A comparison of foot/ground interaction during stair negotiation and level walking in young and older women

Stair design and environmental conditions may play a role in slip accidents on stairs in the workplace, but little is known about the slip resistance requirements on stairs compared to level walking. Older adults have an increased risk of falling compared to younger adults and may be at greater risk during stair negotiation. The purpose of this study was to determine whether the ground reaction force profiles and peak required coefficient of friction (RCOF) differed between young and elderly women or between stair ascent, stair descent and overground walking. While there was a trend towards less risky stair descent behaviour in the older women in terms of their peak RCOF values during stair descent, the increased vertical loading rate in the older women may imply reduced dynamic balance control. The largest mean RCOF peaks occurred during stair ascent in both young and older women, but there were several overground walking trials in both groups and a few stair descent trials of the young women, which resulted in RCOF peaks greater than 0.5. These results should be considered when choosing stair surface materials, particularly in occupational and outdoor settings where the tread surfaces may become wet or contaminated.     

基于足底压力和关节运动信息的步态检测系统设计

作为老年人高发疾病,脑卒中有着较高的几率造成患者行动障碍。及时检测、获取偏瘫患者行走时步态信息,是治疗师为患者制定康复计划的重要环节。目前,较新的步态检测系统大多采用三维分析,三维步态分析系统体积较大、成本高、操作不便,在患者主动康复训练中存在较大的局限。针对现今步态检测系统的局限,设计了一种结合无线传输技术,检测患者关节运动角度及足底压力的运动信息检测系统。系统采用薄膜压力传感器、惯性传感器分别采集足底压力及关节运动信息,实现了对患者的8路足底压力信息及下肢关节运动角度信息的快速、有效采集。实验表明,系统可以有效采集关节运动角度及足底压力信息,在未来的康复训练中有着很好的应用前景。     

Kinematics based sensory fusion for wearable motion assessment in human walking

Measuring the kinematic parameters in unconstrained human motion is becoming crucial for providing feedback information in wearable robotics and sports monitoring. This paper presents a novel sensory fusion algorithm for assessing the orientations of human body segments in long-term human walking based on signals from wearable sensors. The basic idea of the proposed algorithm is to constantly fuse the measured segment's angular velocity and linear acceleration via known kinematic relations between segments. The wearable sensory system incorporates seven inertial measurement units attached to the human body segments and two instrumented shoe insoles. The proposed system was experimentally validated in a long-term walking on a treadmill and on a polygon with stairs simulating different activities in everyday life. The outputs were compared to the reference parameters measured by a stationary optical system. Results show accurate joint angle measurements (error median below 5°) in all evaluated walking conditions with no expressed drift over time.     

Intent Pattern Recognition of Lower-limb Motion Based on Mechanical Sensors

Based on the regularity nature of lower-limb motion, an intent pattern recognition approach for above-knee prosthesis is proposed in this paper. To remedy the defects of recognizer based on electromyogram (EMG), we develop a pure mechanical sensor architecture for intent pattern recognition of lower-limb motion. The sensor system is composed of an accelerometer, a gyroscope mounted on the prosthetic socket, and two pressure sensors mounted under the sole. To compensate the delay in the control of prosthesis, the signals in the stance phase are used to predict the terrain and speed in the swing phase. Specifically, the intent pattern recognizer utilizes intraclass correlation coefficient (ICC) according to the Cartesian product of walking speed and terrain. Moreover, the sensor data are fused via DempsterShafer's theory. And hidden Markov model (HMM) is used to recognize the realtime motion state with the reference of the prior step. The proposed method can infer the prosthesis user's intent of walking on different terrain, which includes level ground, stair ascent, stair descent, up and down ramp. The experiments demonstrate that the intent pattern recognizer is capable of identifying five typical terrain-modes with the rate of 95.8%. The outcome of this investigation is expected to substantially improve the control performance of powered above-knee prosthesis.      

基于生成对抗网络的跨视角步态特征提取

步态是一种能够在远距离、非侵犯的条件下识别身份的生物特征,但在实际场景中,步态很容易受到拍摄视角、行走环境、物体遮挡、着装等因素的影响.在跨视角识别问题上,现有方法只注重将多种视角的步态模板转化到固定视角下,且视角跨度的增大加深了错误的累积.为了提取有效的步态特征用于跨视角步态识别,本文提出了一种基于生成对抗网络的跨视角步态特征提取方法,该方法只需训练一个模型即可将步态模板转换到任意视角下的正常行走状态,并最大化地保留原本的身份特征信息,从而提高步态识别的准确率.在CASIA-B和OUMVLP数据集上的实验结果表明,该方法在解决跨视角步态识别问题上具有一定的鲁棒性和可行性.