柔性下肢外骨骼的设计与信号检测

下肢外骨骼作为一种助力设备,自问世以来一直备受关注.外骨骼的应用不仅有利于解决老龄化人口行动不便的问题,对健康人群而言,也能依靠外骨骼的助力作用减少劳动消耗,减轻劳动负担.设计了基于气动人工肌肉的柔性下肢外骨骼,提升外骨骼助力行为的自适应性与助力效率,减轻外骨骼整体质量,提高穿戴舒适性.实测行走试验分析表明,研制的柔性下肢外骨骼助力效果明确,相比于电机驱动、液压驱动方式,外骨骼的穿戴舒适性显著提升.     

人体步态数据测量系统的设计与实现

下肢外骨骼机器人和下肢康复机器人是当前的研究热点,人体运动的步态数据对其完成步态识别与控制有重要意义.为了获取人体行走过程中的步态信息,本文设计并实现了一套测量足底压力与关节角度数据的数据采集系统.该系统包含微处理器、足底压力传感器、角度传感器、信号调理模块以及数据存储模块,能实现步态数据的自动解算、采集和存储.利用该系统成功采集了不同实验对象在不同速率下足底压力数据和关节的角度数据,结果显示出一定的步态特性,为进一步的步态分析与识别提供参考.     

基于机器学习的人体步态检测智能识别算法研究

为实现快速步态状态判断,以更好地对下肢外骨骼进行高精度的步态识别和控制,进行了基于可穿戴惯性测量装置检测人体姿态变化的算法研究。通过对人体下肢的跌倒、转弯、蹲坐与起立等非周期性步态变化活动进行测算试验,获得了受试者实验过程中身体角度、下肢关节角速度和加速度变化等数据,随后应用随机森林等4种机器学习经典分类算法对受试者进行了活动识别对比分析,结果表明,决策树监督学习算法相对于其他算法,能够快速、准确地检测并判断出人体非周期性变化中的多种活动状态,历次识别精度均可达到99%以上,为可穿戴智能装备的开发与应用提供理论基础。     

时频广义 S 变换和 VL-MOBP 神经网络在人体动作识别中的应用

针对仿生假肢动作识别问题,提出基于时频广义 S 变换和 VL-MOBP 神经网络的下肢动作识别方法。 首先用时频广义 S 变换对年龄在 20~ 40 岁,身高在 170~ 185 cm,体重在 50~ 75 kg 的 22 名男性测试者下肢 4 种表面肌电信号和膝盖弯曲度信号 进行多分辨率分析,得到在时间和频率分辨率较好情况下信号时频累计特性曲线,然后提取时频累计特性曲线幅值的均值和标 准差作为特征向量,用 VL-MOBP 神经网络对人体下肢的行走、站立及静坐 3 种动作进行识别。 实验结果表明,提出的下肢动作 识别方法能够取得很好的识别效果,平均识别准确度达 96. 67%,高出小波变换约 56%,高出短时傅里叶变换约 36%,验证了该 方法在动作识别中的有效性。     

面向康复训练的多通道mRMR-PSO肌电特征选择算法

表面肌电信号的产生超前于肢体运动的发生,具有预测肢体运动的能力,常常辅助患者的康复训练。针对单通道表面肌电信号难以有效进行关节角度预测的问题,本文提出了一种基于多通道肌电特征采集下的最大相关最小冗余结合粒子群优化的特征选择算法。通过与采用mRMR算法、主成分分析法对关节角度预测精度进行实验对比,验证mRMR-PSO算法的性能。实验结果表明,基于多通道的mRMR-PSO特征选择算法相比mRMR和PCA算法在关节角度预测精度分别提高了32.6%和14.9%,从而验证了算法的有效性,并将该算法应用于实际场景。     

按摩筋膜枪，释放你的酸痛触感

肌肉筋膜按摩枪,高频振动唤醒人体机能,释放肌肉酸痛,舒缓肌肉疲劳,拒绝疲劳感,重返好状态。筋膜枪是近一两年开始在大众健身中走红的放松工具,它是一种高频次冲击式的放松方式,通过不同频率对肌肉的冲击达到放松肌肉的效果。筋膜枪在高频次振荡冲击肌肉时会产生一定压力,肌肉张力便会增加,从而激活存在肌腱位置的张力变化感受器—高尔基腱器官,而高尔基腱器具有显著地降低肌肉张力的作用.     

基于肌肉协同的临床步态功能定量评定方法

步态功能评估是脑卒中运动功能评估中的重要组成部分。 为解决临床步态功能评估主观性大,主要依据运动学结果而 非神经肌肉变化进行评估的问题,本文提出基于表面肌电(surface electromyography,sEMG)信号的肌肉协同(muscle synergy)分 析方法对步态功能进行定量评定。 使用非负矩阵分解( non-negative matrix factorization,NMF)算法对预处理后的多通道 sEMG 信号进行分解,计算不同组别不同任务下的协同结构向量相似性,得到协同稳定指数( synergy stability index,SSI)对患者步态功 能进行评估。 组间比较结果表明,测试腿在内侧转弯时,健康对照与患者健侧腿之间的 SSI 差异显著( p = 0. 010),与患者患侧 腿之间的 SSI 也存在显著差异(p = 0. 007);测试腿在外侧转弯时,健康对照与患者患侧腿之间的 SSI 差异显著( p = 0. 036);任务 间比较结果表明,患者健侧腿在外侧转弯时的 SSI 显著高于内侧转弯(p = 0. 017)。 且外侧转弯时患者患侧腿的 SSI 与临床运动 功能量表下肢部分 FMA_LE 显著相关(r = 0. 671,r 2 = 0. 451,p = 0. 033)。 由结果可知,转弯任务下的 SSI 可为临床步态功能的评 估提供一种客观的分析手段,从神经肌肉角度为定量评估提供一种新的思路。     

基于GA-SVR模型的肌力康复电刺激系统的设计研究

为了实现康复电刺激系统治疗参数的个性化定制及实时调整,提出了一种基于调制中频电刺激的下肢肌力康复闭环电刺激系统。设计低频调制中频刺激电路,基于遗传算法建立了电刺激参数与膝关节角度之间的支持向量机回归预测模型,并搭建基于模糊内模控制PID的闭环反馈系统,以达到更精确稳定的参数设置效果。通过膝关节运动实验表明,被试者在无痛感的前提下更接近预期的关节运动轨迹,30组膝关节运动角度与预期值最大均方根误差为10.21°,最小均方根误差为5.48°。相比传统低频电刺激,肌电平均振幅具有20μV以上提升。本文提出的电刺激系统参数可实现因人而异,且可根据闭环反馈结果进行实时调整,该系统能有效活化肌肉、提升肌力,在肌力康复步态训练中有较好的应用前景。     

基于轨迹测量与人机映射的六自由度机械臂运动追踪模型

本研究提出了一种基于六自由度机械臂的遥操作人机系统,旨在设计一种不依赖穿戴设备且直观易用的操控方式。该系统使用KinectV1摄像头及UR3机械臂,以Microsoft骨骼识别库作为基本的人体姿态识别方法,通过人体手臂与机械臂关节的映射,实现机械臂实时追踪人体手臂动作的任务。同时,采用非线性模型预测控制(NMPC)算法对机械臂运动控制进行优化,并设定模糊规则来实现NMPC参数的自适应调整。实验结果表明,在NMPC的优化作用下,机械臂在x及z两个平动方向的平均位移误差和3个转动方向的平均旋转误差以及关节角变化量平均误差都有了显著的降低。测试结果也表明,机械臂整体动作跟随效果良好,验证了本文提出的映射规则和运动学模型的准确性,以及模糊NMPC控制器的有效性。     

基于骨架模型的人体行为分析

随着深度学习运用到图像领域,姿态估计、行为分析等算法的性能得到显著提升,希望在利用较好模型基础上进一步分析,在尽可能短的时间内得到更直观的结果。2016年提出的沙漏堆网络对人体关节点进行多尺度、多阶段的训练,在MPII数据集上回归了16对关节点坐标,在单个11 G显存的GPU上的平均准确率为87.6%;连接关节点构建人体骨架模型,然后根据骨架模型的加权角和倾斜角等几何特征,进一步推断人体的动作和行为状态,最后对人体行为进行分类和判断,包括站立、直坐、躺下等常见7类动作,平均准确率为82%,优势在于有效降低计算量和处理时间。     

Medial Gastrocnemius Myoelectric Control of a Robotic Ankle Exoskeleton

A previous study from our laboratory showed that when soleus electromyography was used to control the amount of plantar flexion assistance from a robotic ankle exoskeleton, subjects significantly reduced their soleus activity to quickly return to normal gait kinematics. We speculated that subjects were primarily responding to the local mechanical assistance of the exoskeleton rather than directly attempting to reduce exoskeleton mechanical power via decreases in soleus activity. To test this observation we studied ten healthy subjects walking on a treadmill at 1.25 m/s while wearing a robotic exoskeleton proportionally controlled by medial gastrocnemius activation. We hypothesized that subjects would primarily decrease soleus activity due to its synergistic mechanics with the exoskeleton. Subjects decreased medial gastrocnemius recruitment by 12% (${ p}≪ 0.05$ ) but decreased soleus recruitment by 27% ( ${ p}≪ 0.05$). In agreement with our hypothesis, the primary reduction in muscle activity was not for the control muscle (medial gastrocnemius) but for the anatomical synergist to the exoskeleton (soleus). These findings indicate that anatomical morphology needs to be considered carefully when designing software and hardware for robotic exoskeletons.      

Assessment of motion of a swing leg and gait rehabilitation with a gravity balancing exoskeleton.

The gravity balancing exoskeleton, designed at University of Delaware, Newark, consists of rigid links, joints and springs, which are adjustable to the geometry and inertia of the leg of a human subject wearing it. This passive exoskeleton does not use any motors but is designed to unload the human leg joints from the gravity load over its range-of-motion. The underlying principle of gravity balancing is to make the potential energy of the combined leg-machine system invariant with configuration of the leg. Additionally, parameters of the exoskeleton can be changed to achieve a prescribed level of gravity assistance, from 0% to 100%. The goal of the results reported in this paper is to provide preliminary quantitative assessment of the changes in kinematics and kinetics of the walking gait when a human subject wears such an exoskeleton. The data on kinematics and kinetics were collected on four healthy and three stroke patients who wore this exoskeleton. These data were computed from the joint encoders and interface torque sensors mounted on the exoskeleton. This exoskeleton was also recently used for a six-week training of a chronic stroke patient, where the gravity assistance was progressively reduced from 100% to 0%. The results show a significant improvement in gait of the stroke patient in terms of range-of-motion of the hip and knee, weight bearing on the hemiparetic leg, and speed of walking. Currently, training studies are underway to assess the long-term effects of such a device on gait rehabilitation of hemiparetic stroke patients.     

Biomechanical evaluation of a prototype foot/ankle prosthesis.

In this paper, we report on our pilot evaluation of a prototype foot/ankle prosthesis. This prototype has been designed and fabricated with the intention of providing decreased ankle joint stiffness during the middle portion of the stance phase of gait, and increased (i.e., more normal) knee range of motion during stance. Our evaluation involved fitting the existing prototype foot/ankle prosthesis, as well as a traditional solid ankle cushioned heel (SACH) foot, to an otherwise healthy volunteer with a below-knee (BK) amputation. We measured this individual's lower extremity joint kinematics and kinetics during walking using a video motion analysis system and force platform. These measurements permitted direct comparison of prosthetic ankle joint stiffness and involved side knee joint motion, as well as prosthetic ankle joint moment and power.     

Simulation of a functional neuromuscular stimulation powered mechanical gait orthosis with coordinated joint locking.

The purpose of this study was to examine a hybrid orthosis system (HOS) for walking after spinal-cord injury (SCI) that coordinates the mechanical locking and unlocking of knee and ankle joints of a reciprocating gait orthosis (RGO), while propulsive forces are injected and unlocked joints controlled with functional neuromuscular stimulation (FNS). The likely effectiveness of the HOS in terms of forward progression, stability, and posture of paraplegic gait was determined in this simulation study. A three-dimensional computer model of a HOS combining FNS with an RGO incorporating feedback control of muscle activation and joint locking was developed. An anthropomorphic human model included passive joint moments and a foot-ground contact model adapted from other studies. A model of the RGO reciprocally coupled the hips and locked and unlocked the knee and ankle joints during stance and swing respectively. The actions of muscles under FNS activation were modeled via closed-loop control of joint torque inputs. A walking aid that mimicked canes and voluntary upper extremity actions maintained lateral stability by providing the necessary shoulder forces and moments. The simulated HOS achieved gait speeds of 0.51 +/- 0.03 m/s, stride lengths of 0.85 +/- 0.04 m, and cadences of 72 +/- 4 steps/min, exceeding the reported performance of other assistive gait systems. Although minimal forward trunk tilt was found to be necessary during specific phases of gait, posture, and stability were significantly improved over FNS-only systems.     

Postural constraints for two forms of quadrupedal walking

Summary form only given. Quadrupeds exhibit several forms of locomotion. One area of study is the demands made on the nervous system to meet the constraints of different forms. Two forms of walking, forward (FWD) and backward (BWD), in cats trained to walk on a treadmill (0.6 m/s) are discussed. It is shown that the basic muscle synergies are the same for both forms: extensors are active during stance and flexors during swing. Although synergies during the step cycle are similar, the timing and EMG amplitude, as well as the limb kinematics, are form specific. During FWD walking, for example, the hip extended during stance while the hip flexed during BWD stance. Thus, single-joint muscles such as the anterior biceps femoris, active during stance, shorten during FWD stance and lengthen during BWD stance, and feedback from hip joint and myotatic receptors are different even though the motor output is similar. It is postulated that BWD walking is possible only when the quadruped is able to set appropriate posture     

The effects of long-term FES-assisted walking on intrinsic and reflex dynamic stiffness in spastic spinal-cord-injured subjects

The effects of long-term functional electrical stimulation (FES)-assisted walking on ankle dynamic stiffness were examined in spinal cord-injured (SCI) subjects with incomplete motor function loss. A parallel-cascade system identification method was used to identify intrinsic and reflex contributions to dynamic ankle stiffness at different ankle positions while subjects remained relaxed. Intrinsic stiffness dynamics were well modeled by a linear second-order model relating intrinsic torque to joint position. Reflex stiffness dynamics were accurately described by a linear third-order model relating halfwave rectified velocity to reflex torque. We examined four SCI subjects before and after long-term FES-assisted walking (>16 mo). Another SCI subject, who used FES for only five months was examined 12 mo latter to serve as a non-FES, SCI control. Reflex stiffness decreased in FES subjects by an average of 53% following FES-assisted walking, intrinsic stiffness also dropped by 45%. In contrast, both reflex and intrinsic stiffness increased in the non-FES, SCI control. These findings suggest that FES-assisted walking may have therapeutic effects, helping to reduce abnormal joint stiffness.     

基于微电机的髋关节辅助外骨骼的研究

针对广大老年人普遍存在的行走不便问题,该文设计了一种以TMS320F2812为核心,基于微电机的便携式髋关节辅助外骨骼系统。首先介绍了髋关节外骨骼系统工作原理、硬件组成及相应软件结构;然后,讨论了电机转速受限状态下,输出转矩的控制模型,采用最小拍控制算法对电机电枢电流进行伺服控制;通过调节电压PWM波占空比实现对电机负载电流的控制,进而完成对电机输出转矩的控制。最终保证了髋关节辅助外骨骼提供的辅助力矩光滑、平稳。     

Robot Assisted Gait Training With Active Leg Exoskeleton (ALEX)

Gait training of stroke survivors is crucial to facilitate neuromuscular plasticity needed for improvements in functional walking ability. Robot assisted gait training (RAGT) was developed for stroke survivors using active leg exoskeleton (ALEX) and a force-field controller, which uses assist-as-needed paradigm for rehabilitation. In this paradigm undesirable gait motion is resisted and assistance is provided towards desired motion. The force-field controller achieves this paradigm by effectively applying forces at the ankle of the subject through actuators on the hip and knee joints. Two stroke survivors participated in a 15-session gait training study each with ALEX. The results show that by the end of the training the gait pattern of the patients improved and became closer to a healthy subject's gait pattern. Improvement is seen as an increase in the size of the patients' gait pattern, increased knee and ankle joint excursions and increase in their walking speeds on the treadmill.      

The Effects on Kinematics and Muscle Activity of Walking in a Robotic Gait Trainer During Zero-Force Control

“Assist as needed” control algorithms promote activity of patients during robotic gait training. Implementing these requires a free walking mode of a device, as unassisted motions should not be hindered. The goal of this study was to assess the normality of walking in the free walking mode of the LOPES gait trainer, an 8 degrees-of-freedom lightweight impedance controlled exoskeleton. Kinematics, gait parameters and muscle activity of walking in a free walking mode in the device were compared with those of walking freely on a treadmill. Average values and variability of the spatio-temporal gait variables showed no or small (relative to cycle-to-cycle variability) changes and the kinematics showed a significant and relevant decrease in knee angle range only. Muscles involved in push off showed a small decrease, whereas muscles involved in acceleration and deceleration of the swing leg showed an increase of their activity. Timing of the activity was mainly unaffected. Most of the observed differences could be ascribed to the inertia of the exoskeleton. Overall, walking with the LOPES resembled free walking, although this required several adaptations in muscle activity. These adaptations are such that we expect that Assist as Needed training can be implemented in LOPES.