Papers by Keyword: Exoskeleton

Abstract: Over the past decade, noninvasive brain-computer interfaces (BCIs) leveraging electroencephalography(EEG) to decode gait intention have matured from proof-of-concept studies to nearclinicalimplementations. We systematically reviewed 55 studies (January 2015-April 2024) usingPRISMA guidelines, focusing on neurophysiological markers (MRCPs, ERD/ERS, high-γ), signalprocessing pipelines (artifact suppression, time-frequency transforms), machine-learning classifiers(CSP-SVM, ERD SVM), and deep-learning frameworks (spatio-spectral CNNs, LSTM RNNs). Acrossstudies, median classification accuracy rose from 75% (2015-2018) to 87% (2021-2024), while detectionlatency fell below 200ms. Innovations include enhancing intention detection with emotionevokingmusic stimuli (up to early ERD and improved accuracy), decoding pediatric gait kinematicswith state-space models (r = 0.71 hip, 0.59 knee), and session-to-session transfer learning withoutrecalibration (≤4% performance drop). Challenges remain in artifact mitigation, small sample sizes,and limited multi-centre trials. We propose open, standardized datasets, transfer-learning pipelines,and larger clinical validations to accelerate translation.

Abstract: The choice of using speech to control the exoskeleton is based on the number of exoskeletons that are controlled using the EMG signal, where the EMG signal itself has the weakness of the complexity of the signal which is influenced by the position of the electrodes, as well as muscle fatigue. The purpose of this research is to develop an exoskeleton device using voice control based on embedded machine learning on a Raspberry Pi minicomputer. In this study, two feature extraction types namely mel-frequency cepstral coefficient (MFCC) and zero-crossing (ZC), and two machine learning algorithms, namely K-nearest Neighbor (K-NN) and Decision Tree (DT) were evaluated. The hand exoskeleton development consists of 3D hand design, microphone, Raspberry Pi 4B+, PCA9685 servo driver, and servo motor. Microphone was used to record voice commands given. After model evaluation, it was found that the MFCC extraction combined with the K-NN algorithm and the best accuracy (96±7.0%). In the implementation, we found that the accuracy is 79±14.46% and 90±14.14% for open and close commands.

Abstract: In this paper experimental investigations concerning the prototype validation of an exoskeleton for human gait rehabilitation are presented. The proposed exoskeleton is intended for human legs motion assistance, being a low cost solution. The exoskeleton provides motion assistance for human hip and knee joint. The experimental gait analysis of the exoskeleton and assisted human subject is performed with goniometers sensors. These sensors are attached to the exoskeleton hip and knee joints and measure the angle variation. The results obtained for the exoskeleton motion are compared with those obtained in case of a healthy human subject. Angular variation corresponding to a step, as a comparison for human and exoskeleton joints is performed. The conclusion is that the designed exoskeleton is proper for human gait motion assistance, because the joints motion is comparable with the human subject.

Abstract: Among the problems of major concern faced by the geriatric community the fore most is considered to be: “as we grow older it gets difficult to walk”. As they lose their strength to withstand their weight they become weak to walk on their own. Through this project a design is proposed for an assistive modular lower limb exoskeleton robot to enable aged people to walk on their own. The design is mainly based on the amplification of the pressure applied on the thighs and ankles of the legs and these pressures are used to move the legs of the robot which support the legs of the user. Since the user is capable of generating a minimum pressure on their own the working of the gait pattern is mainly based on the movement of knee and ankle, and so only the knee and the ankle are considered in this design. Another reason is that the major weight of the body acts on the knee and the foot while walking. The working of the robot is mainly based on the layered control algorithm embedded on the microcontroller acting through the sensors and actuators. Pressure sensors are used to measure the applied pressure and electrical actuators are used due to their lesser weight. Based on the experimental results obtained with the working robot the design would be fine-tuned for optimized performance.

Abstract: Exoskeleton is an external cantilever to support inner living organism organs. Some animals have the exoskeleton attached naturally on their body as those found in a million species of arthropods order. The word “Arthropods” itself are related to their body plan—their segmented body, hard exoskeleton, and jointed appendages [1]. This structure inspired us to implement the exoskeleton to help humans who are permanently or temporarily have problems with their lower limb, for example those who are recovering from stroke or road-accidents. Stroke is a condition where lower limb is partially paralyzed (Hemiplegia) and causes inability to walk normally or to do other physical activities [2].The exoskeleton consists of four main parts; frame for the leg (XOTF), frame for the thigh (XOF), the belt, and the controller stick. The frame is designed using lightweight material. Considering the material’s availability and price at the local market, we choose Aluminum Composite Panel (ACP) as frame’s material. The belt acts as the brain for the exoskeleton, where it provides electrical power and the source of motion. There are four frames on the exoskeleton which moves simultaneously while walking. In the design, we use two joy-sticks as main controller to adjust position of each exoskeleton frame. The sticks are attached to standing-canes which help the user to improve stability. The sticks have buttons and lever to control the exoskeleton and adjust thighs and legs fit for several moves; stand, sit, walk, going up and downstairs, or any other position which need balance and power.The exoskeleton design has been tested for its motion and stability. The results show that the design is suitable for those who are fully or partially paralyzed on their lower limb.

Abstract: Based on upper limb’s biomechanisms, in this paper, a robotic rehabilitation system is presented. It is designed as a 4 DOFs wearable exoskeleton applicable for repetitive practice of passive or active movements of the arm in shoulder joint and forearm in elbow joint. The kinematic analysis of the proposed system is followed by the 3D model and a description of the developed prototype.

Abstract: The design of lower limb rehabilitation robot can be categorized into two approaches: the end-effector and the exoskeleton. Both types of the robots have different advantages and disadvantages. The exoskeleton type is designed to mimic the kinematic structure of the human skeleton by controlling hip and knee joints but the end-effector type is driven at the footplate which allows patients to perform various gait training exercise. In this paper, the end-effector and exoskeleton device are compared based on dynamical analysis using Matlab's Simechanics simulation. The hybrid lower limb rehabilitation robot is also proposed based on the exoskeleton robot with the adjustable mechanical coupling interface between human and robot and the active footplate. The hybrid design combines the advantages of both the exoskeleton and the end-effector by allowing the mechanical coupling parameters and the active footplate controller to be adjustable at different stages of training. The proposed design can improve both joints misalignment and joint trajectory tracking problems in both existing approaches.

Abstract: Functional recovery of upper limb after stroke is crucial to restore the ability to perform activities of daily living (ADL). This paper presents a robotic rehabilitation approach based on repetitive exercise aimed to help stroke survivors relearn the skills of finger flexion and extension at the comfort of their home. The finger rehabilitation device deploys Shape Memory Alloy (SMA) wires as an actuation approach to deliver three degrees of freedom per finger module. The advantages and challenges of using SMA wires rather than conventional actuators are discussed. A prototype of the finger rehabilitation device was built using PLA material and experiments have been conducted for the purposes of feasibility study. Tests conducted on the wires suggest that it has to have sufficient weight and also has to be stretching at high temperature rather than room temperature in order to have an optimum range of recovery.

Abstract: A current study regarding the development of a rehabilitation device for lower extremity is presented. The device is specifically designed for rehabilitation of post-stroke patients who encounter walking weakness. The rehabilitation device is categorized as an active device that is power-driven by a DC motor. Its design was the outcome of improvements to counter the problems existed in the a newly developed prototype. A motion simulation was used to ratify the motion capability of the proposed design in the modelling design process. Analysis of kinematic and dynamic behavior of the motion simulation has been carried out. The smoothness of the mechanical movement and the linear velocity provided by the device are acceptable as additional work of the design process. Based on a statistical analysis, the study found that there is no significant difference from the motor torque requirement even though the angular velocity of the motor was changed substantially.

Abstract: This paper presents design and application of foot modules in exoskeleton. Detecting the walking phase and intent is most important data for controlling the wearable exoskeleton. The zero moment point (ZMP) trajectory in the robot foot support area is significant criterion for a stability of the walk. The ZMP is calculated by measured force / torque data and Force / Torque sensor (F/T sensor) can measure that. But, the thick foot module is inconvenient to walking. For such a reason, most foot modules are developed using the film type sensors such as force sensing resistor (FSR), ground contact sensor such as tape switch and so on. In this paper, the foot modules equipped optimal design of 3-axis F/T sensor and ground contact sensor are presented and experimental application of detecting the walking phase using both sensors is introduced.