scholarly journals An FPGA-Based Upper-Limb Rehabilitation Device for Gesture Recognition and Motion Evaluation Using Multi-Task Recurrent Neural Networks

2021 ◽  
Author(s):  
Haoyan Liu ◽  
Atiyehsadat Panahi ◽  
David Andrews ◽  
Alexander Nelson
Keyword(s):  
Neural Networks ◽  
Upper Extremity ◽  
Upper Limb ◽  
Recurrent Neural Networks ◽  
Spinal Cord Injuries ◽  
Low Cost ◽  
Motor Impairment ◽  
Current Therapy ◽  
Upper Limb Rehabilitation ◽  
Machine Learning Classification

Upper-Extremity motor impairment affects millions of Americans due to cerebrovascular incidents, spinal cord injuries, or brain trauma. Current therapy practices used to assist these individuals in regaining motor functionality often require extensive time at rehabilitation facilities with potentially prohibitive travel or financial costs. This work presents a mobile low-cost field programmable gate array (FPGA)-smart rehabilitation system that can be used in home environments. The prototype is a rehabilitation table instrumented with a capacitive sensor array (CSA) to track upper-extremity motions of the user through proximity or touch. In addition, inertial measurement units (IMUs) are placed on the affected upper limb and combined with the CSA data with our sensor fusion signal processing architecture. Motions are classified and evaluated using multi-task convolutional recurrent neural networks with three additional motion quality output classes to personalize recognition based on the particular motor skills of each patient. The prototype achieves above 99% accuracy with 32-bit fixed-point format implementation for recognizing dynamic motions and identifying unnatural characteristics (i.e., tremor or limited flexion and extension) in upper limb motions based on sensor values. The convolutional recurrent neural network (C-RNN) fusion classification network is implemented on a 200 MHz Zynq ZCU104 FPGA using an HLS-based design optimized with pipelining and parallelism techniques and achieves 5.4x speedup compared to ARM® Cortex-A53 implementation running at an operating frequency of 1.3 GHz. The prototype is also demonstrated to perform the machine learning classification in real-time.

scholarly journals An FPGA-Based Upper-Limb Rehabilitation Device for Gesture Recognition and Motion Evaluation Using Multi-Task Recurrent Neural Networks

2021 ◽  
Author(s):  
Haoyan Liu ◽  
Atiyehsadat Panahi ◽  
David Andrews ◽  
Alexander Nelson
Keyword(s):  
Neural Networks ◽  
Upper Extremity ◽  
Upper Limb ◽  
Recurrent Neural Networks ◽  
Spinal Cord Injuries ◽  
Low Cost ◽  
Motor Impairment ◽  
Current Therapy ◽  
Upper Limb Rehabilitation ◽  
Machine Learning Classification

Upper-Extremity motor impairment affects millions of Americans due to cerebrovascular incidents, spinal cord injuries, or brain trauma. Current therapy practices used to assist these individuals in regaining motor functionality often require extensive time at rehabilitation facilities with potentially prohibitive travel or financial costs. This work presents a mobile low-cost field programmable gate array (FPGA)-smart rehabilitation system that can be used in home environments. The prototype is a rehabilitation table instrumented with a capacitive sensor array (CSA) to track upper-extremity motions of the user through proximity or touch. In addition, inertial measurement units (IMUs) are placed on the affected upper limb and combined with the CSA data with our sensor fusion signal processing architecture. Motions are classified and evaluated using multi-task convolutional recurrent neural networks with three additional motion quality output classes to personalize recognition based on the particular motor skills of each patient. The prototype achieves above 99% accuracy with 32-bit fixed-point format implementation for recognizing dynamic motions and identifying unnatural characteristics (i.e., tremor or limited flexion and extension) in upper limb motions based on sensor values. The convolutional recurrent neural network (C-RNN) fusion classification network is implemented on a 200 MHz Zynq ZCU104 FPGA using an HLS-based design optimized with pipelining and parallelism techniques and achieves 5.4x speedup compared to ARM® Cortex-A53 implementation running at an operating frequency of 1.3 GHz. The prototype is also demonstrated to perform the machine learning classification in real-time.

Design and Kinematic Analysis of Co-Exoskeleton with Passive Translational Joints for Upper-Limb Rehabilitation

2018 ◽  
Vol 15 (05) ◽  
pp. 1850020 ◽  
Author(s):  
Leiyu Zhang ◽  
Jianfeng Li ◽  
Junhui Liu ◽  
Peng Su ◽  
Chunzhao Zhang
Keyword(s):  
Upper Limb ◽  
Kinematic Analysis ◽  
Compensation Effect ◽  
Spinal Cord Injuries ◽  
Glenohumeral Joint ◽  
Motor Impairment ◽  
Vertical Movement ◽  
Upper Limb Rehabilitation ◽  
Kinematic Models ◽  
Limb Rehabilitation

A key approach for reducing motor impairment and regaining independence after spinal cord injuries or strokes is frequent and repetitive functional training. A compatible exoskeleton (Co-Exoskeleton) with four passive translational joints is proposed for the upper-limb rehabilitation. There are only three passive translational joints to track and assist movements of the glenohumeral joint (GH), where two joints are installed horizontally at the front section and another one at the connecting interface of the upper arm. This type of configuration can lower the influences of gravities of the exoskeleton device and upper extremity. The kinematic models of GH and the corresponding human–machine system are established using the analytical method. A numerical simulation of the kinematic models is implemented with MATLAB to emphatically analyze the kinematic characteristics of passive joints and the center of Co-Exoskeleton. The translational displacements of passive joints in four elevation planes are obtained during the elevating process. The results of the kinematic analysis show that the passive joints have similar motion characteristics under different elevation planes. Additionally, the position changes of GH in three directions can be tracked and compensated approximately. Co-Exoskeleton has an especially good compensation effect for the vertical movement of GH. The compensation effect and kinematic models are verified by using the elevating experiments. This research provides theoretical and methodological guidance for the ergonomic design and kinematic analysis of the rehabilitation exoskeleton.

scholarly journals Binary and Multiclass Text Classification by Means of Separable Convolutional Neural Network

Inventions ◽  
2021 ◽  
Vol 6 (4) ◽  
pp. 70
Author(s):  
Elena Solovyeva ◽  
Ali Abdullah
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Convolutional Neural Network ◽  
Recurrent Neural Networks ◽  
Low Cost ◽  
Computational Cost ◽  
High Accuracy ◽  
Activation Functions ◽  
Fully Connected ◽  
Fully Connected Networks

In this paper, the structure of a separable convolutional neural network that consists of an embedding layer, separable convolutional layers, convolutional layer and global average pooling is represented for binary and multiclass text classifications. The advantage of the proposed structure is the absence of multiple fully connected layers, which is used to increase the classification accuracy but raises the computational cost. The combination of low-cost separable convolutional layers and a convolutional layer is proposed to gain high accuracy and, simultaneously, to reduce the complexity of neural classifiers. Advantages are demonstrated at binary and multiclass classifications of written texts by means of the proposed networks under the sigmoid and Softmax activation functions in convolutional layer. At binary and multiclass classifications, the accuracy obtained by separable convolutional neural networks is higher in comparison with some investigated types of recurrent neural networks and fully connected networks.

scholarly journals Diseño mecánico de un exoesqueleto para rehabilitación de miembro superior

2015 ◽  
Vol 17 (1) ◽  
pp. 79-90 ◽  
Author(s):  
Juan Francisco Ayala Lozano ◽  
Guillermo Urriolagoitia Sosa ◽  
Beatriz Romero Ángeles ◽  
Christopher René Torres San-Miguel ◽  
Luis Antonio Aguilar-Pérez ◽  
...  
Keyword(s):  
Upper Limb ◽  
Degrees Of Freedom ◽  
Mechanical Design ◽  
Low Cost ◽  
Structure Design ◽  
Upper Limb Rehabilitation ◽  
Palabras Clave ◽  
Mexican Patient ◽  
Almost All ◽  
Limb Rehabilitation

<strong>Título en ingles: Mechanical design of an exoskeleton for upper limb rehabilitation</strong><p><strong>Título corto: Diseño mecánico de un exoesqueleto</strong></p><p><strong>Resumen:</strong> El ritmo de vida actual, tanto sociocultural como tecnológico, ha desembocado en un aumento de enfermedades y padecimientos que afectan las capacidades físico-motrices de los individuos. Esto ha originado el desarrollo de prototipos para auxiliar al paciente a recuperar la movilidad y la fortaleza de las extremidades superiores afectadas. El presente trabajo aborda el diseño de una estructura mecánica de un exoesqueleto con 4 grados de libertad para miembro superior. La cual tiene como principales atributos la capacidad de ajustarse a la antropometría del paciente mexicano (longitud del brazo, extensión del antebrazo, condiciones geométricas de la espalda y altura del paciente). Se aplicó el método <em>BLITZ QFD</em> para obtener el diseño conceptual óptimo y establecer adecuadamente las condiciones de carga de servicio. Por lo que, se definieron 5 casos de estudio cuasi-estáticos e implantaron condiciones para rehabilitación de los pacientes. Asimismo, mediante el Método de Elemento Finito (MEF) se analizaron los esfuerzos y deformaciones a los que la estructura está sometida durante la aplicación de los agentes externos de servicio. Los resultados presentados en éste trabajo exhiben una nueva propuesta para la rehabilitación de pacientes con problemas de movilidad en miembro superior. Donde el equipo propuesto permite la rehabilitación del miembro superior apoyado en 4 grados de libertad (tres grados de libertad en el hombro y uno en el codo), el cual es adecuado para realizar terapias activas y pasivas. Asimismo, es un dispositivo que está al alcance de un mayor porcentaje de la población por su bajo costo y fácil desarrollo en la fabricación.</p><p><strong>Palabras clave:</strong> MEF, Blitz QFD, exoesqueletos, diseño mecánico.</p><p><strong>Abstract</strong>: The pace of modern life, both socio-cultural and technologically, has led to an increase of diseases and conditions that affect the physical-motor capabilities of persons. This increase has originated the development of prototypes to help patients to regain mobility and strength of the affected upper limb. This work, deals with the mechanical structure design of an exoskeleton with 4 degrees freedom for upper limb. Which has the capacity to adjust to the Mexican patient anthropometry (arm length, forearm extension, geometry conditions of the back and the patient’s height) BLITZ QFD method was applied to establish the conceptual design and loading service conditions on the structure.  So, 5 quasi-static cases of study were defined and conditions for patient rehabilitation were subjected. Also by applying the finite element method the structure was analyzed due to service loading. The results presented in this work, show a new method for patient rehabilitation with mobility deficiencies in the upper limb. The proposed new design allows the rehabilitation of the upper limb under 4 degrees of freedom (tree degrees of freedom at shoulder and one at the elbow), which is perfect to perform active and passive therapy. Additionally, it is an equipment of low cost, which can be affordable to almost all the country population.</p><p><strong>Key words:</strong> FEM, Blitz QFD, exoskeletons, mechanical design<strong>.</strong></p><p><strong>Recibido:</strong> agosto 20 de 2014   <strong>Aprobado:</strong> marzo 26 de 2015</p>

Parallel Articulated-Cable Exercise Robot (PACER): Novel Home-Based Cable-Driven Parallel Platform Robot for Upper Limb Neuro-Rehabilitation

2015 ◽  
Author(s):  
Aliakbar Alamdari ◽  
Venkat Krovi
Keyword(s):  
Upper Limb ◽  
Degrees Of Freedom ◽  
Dimensional Space ◽  
Low Cost ◽  
Three Dimensional ◽  
Parallel Architecture ◽  
Upper Limb Rehabilitation ◽  
Home Based ◽  
Tensile Forces ◽  
Parallel Platform

This paper examines the design, analysis and control of a novel hybrid articulated-cable parallel platform for upper limb rehabilitation in three dimensional space. The proposed lightweight, low-cost, modular reconfigurable parallel-architecture robotic device is comprised of five cables and a single linear actuator which connects a six degrees-of-freedom moving platform to a fixed base. This novel design provides an attractive architecture for implementation of a home-based rehabilitation device as an alternative to bulky and expensive serial robots. The manuscript first examines the kinematic analysis prior to developing the dynamic equations via the Newton-Euler formulation. Subsequently, different spatial motion trajectories are prescribed for rehabilitation of subjects with arm disabilities. A low-level trajectory tracking controller is developed to achieve the desired trajectory performance while ensuing that the unidirectional tensile forces in the cables are maintained. This is now evaluated via a simulation case-study and the development of a physical testbed is underway.

scholarly journals Upper Limb Rehabilitation Robot Powered by PAMs Cooperates with FES Arrays to Realize Reach-to-Grasp Trainings

2017 ◽  
Vol 2017 ◽  
pp. 1-15 ◽  
Author(s):  
Xikai Tu ◽  
Hualin Han ◽  
Jian Huang ◽  
Jian Li ◽  
Chen Su ◽  
...  
Keyword(s):  
Upper Limb ◽  
Low Cost ◽  
Hand Function ◽  
Rehabilitation Robot ◽  
Reach To Grasp ◽  
Daily Lives ◽  
Upper Limb Rehabilitation ◽  
Highly Nonlinear ◽  
Rehabilitation System ◽  
Limb Rehabilitation

The reach-to-grasp activities play an important role in our daily lives. The developed RUPERT for stroke patients with high stiffness in arm flexor muscles is a low-cost lightweight portable exoskeleton rehabilitation robot whose joints are unidirectionally actuated by pneumatic artificial muscles (PAMs). In order to expand the useful range of RUPERT especially for patients with flaccid paralysis, functional electrical stimulation (FES) is taken to activate paralyzed arm muscles. As both the exoskeleton robot driven by PAMs and the neuromuscular skeletal system under FES possess the highly nonlinear and time-varying characteristics, iterative learning control (ILC) is studied and is taken to control this newly designed hybrid rehabilitation system for reaching trainings. Hand function rehabilitation refers to grasping. Because of tiny finger muscles, grasping and releasing are realized by FES array electrodes and matrix scan method. By using the surface electromyography (EMG) technique, the subject’s active intent is identified. The upper limb rehabilitation robot powered by PAMs cooperates with FES arrays to realize active reach-to-grasp trainings, which was verified through experiments.

scholarly journals Multi-Domain Dynamic Modelling of a Low-Cost Upper Limb Rehabilitation Robot

Robotics ◽  
2021 ◽  
Vol 10 (4) ◽  
pp. 134
Author(s):  
Adam G. Metcalf ◽  
Justin F. Gallagher ◽  
Andrew E. Jackson ◽  
Martin C. Levesley
Keyword(s):  
Dynamic Model ◽  
Upper Limb ◽  
Low Cost ◽  
Equations Of Motion ◽  
Dynamic Modelling ◽  
Rehabilitation Robot ◽  
Data Logging ◽  
Upper Limb Rehabilitation ◽  
Model Match ◽  
Limb Rehabilitation

Tracking patient progress through a course of robotic tele-rehabilitation requires constant position data logging and comparison, alongside periodic testing with no powered assistance. The test data must be compared with previous test attempts and an ideal baseline, for which a good understanding of the dynamics of the robot is required. The traditional dynamic modelling techniques for serial chain robotics, which involve forming and solving equations of motion, do not adequately describe the multi-domain phenomena that affect the movement of the rehabilitation robot. In this study, a multi-domain dynamic model for an upper limb rehabilitation robot is described. The model, built using a combination of MATLAB, SimScape, and SimScape Multibody, comprises the mechanical electro-mechanical and control domains. The performance of the model was validated against the performance of the robot when unloaded and when loaded with a human arm proxy. It is shown that this combination of software is appropriate for building a dynamic model of the robot and provides advantages over the traditional modelling approach. It is demonstrated that the responses of the model match the responses of the robot with acceptable accuracy, though the inability to model backlash was a limitation.

Sign in / Sign up

Export Citation Format

Share Document