Back-support exoskeletons may reduce physical loading during lifting and stooping, but their effects are usually assessed indirectly using EMG and motion capture because individual muscle forces cannot be measured in vivo. This thesis developed and evaluated a proof-of-concept OpenSim Moco musculoskeletal modelling predictive simulation framework to estimate how a passive back-support exoskeleton affects muscle forces during stooping.

Experimental data from nine healthy participants performing stoop-lifting trials with and without the passive rigid Laevo exoskeleton included synchronised EMG and motion capture. Analyses examined erector spinae activation, marker-based movement proxies, and hip and leg muscle activation. To determine significance, paired-sample t-tests were used. In OpenSim, the Laevo was modelled as a passive angle-dependent hip flexion torsional spring.

The experimental results did not show a statistically significant reduction in erector spinae activation when using the Laevo, for either peak activation (p = 0.948) or average activation (p = 0.370). The shoulder-hip distance results suggested a trend towards a smaller motion range when using the Laevo, especially in the empty-crate condition (p = 0.090), but this did not reach statistical significance. Clear statistical evidence for a general change in hip or leg muscle activation was not found, although average biceps femoris activation was significantly lower with the Laevo, decreasing from 0.563 to 0.484 in normalized activation (p = 0.016). The model showed partial qualitative agreement with some experimental trends, especially for selected movement-pattern and muscle-activation outcomes.

Overall, the developed framework generated predictive simulations with and without exoskeleton support. However, the results should be interpreted as a preliminary proof of concept rather than model validation. The main contribution is a reproducible OpenSim Moco workflow for modelling a passive back-support exoskeleton and comparing it with experimental EMG and motion-capture trends.
... Read more

Intuitive control of assistive robotic devices, such as exoskeletons and arm supports, requires inferring the user’s interaction with objects in the environment. Surface electromyography (EMG) and inertial measurement units (IMU) provide complementary information about muscle activation and limb kinematics, but interpreting these sensory modalities for real-time control remains challenging. Deep learning is effective for modeling human motion intention, but has seen limited use in estimating the handheld load during object manipulation. This paper proposes a sensor-fused spatio-temporal transformer (ST-Transformer) that regresses the handheld load from synchronized EMG and IMU signals, together with a real-time acquisition and processing pipeline for an arm support device. Data were used from 17 participants performing a weight-movement task spanning six weight classes (0−6kg). EMG and IMU normalization, dataset-balancing augmentation, dropout, and weight decay were applied to improve cross-participant generalization. Trained and tested on the same participants, the sensor-fused model estimated load accurately (all metrics participant-class-balanced; R2 =0.935, MAE = 0.316kg, RMSE = 0.441kg) and significantly outperformed an EMG-only model (R2 = 0.913, MAE = 0.380kg, RMSE = 0.520kg). Under Leave-One-Participant-Out (LOPO) cross-validation, however, the fused model (R2 = 0.853, MAE = 0.536kg, RMSE = 0.680kg) retained only a slight, statistically non-significant edge over EMG alone (R2 =0.839, MAE =0.546kg, RMSE =0.703kg), while the IMU-only model degraded sharply. This indicates that the transferable load information is carried primarily by muscle activation, while the complementary IMU contribution is largely entangled with participant-specific characteristics. An attribution analysis localizes the load-relevant signal to the forearm muscles, indicating that a compact forearm-worn sensor set captures most of the usable signal, and the model (approximately 1.03 106 parameters) is feasible for real-time on-device inference on current microcontrollers. ... Read more

Stroke is a leading cause of long-term disability and frequently results in upper-limb motor impairments that reduce manual dexterity and quality of life. Robotic assessment systems have been developed to provide objective quantification of these impairments; however, their high cost limits accessibility, particularly in low- and middle-income countries where the majority of stroke cases occur. This thesis investigates the feasibility of implementing admittance control in an ultra-low-cost haptic device intended as a hand-module extension for robotic stroke assessment. A single-degree-of-freedom haptic paddle was developed using widely available components and a total hardware budget below €50. The device incorporates force and position sensing, a DC motor actuation system, and an Arduino-based control architecture. An admittance controller was implemented to render programmable virtual inertia, damping, and stiffness, while a PID position controller provided motion tracking. System performance was evaluated through step-response, frequency-response, and position-tracking experiments involving human interaction. The developed device successfully achieved stable admittance-controlled interaction and is capable of rendering a range of virtual dynamic environments. Experimental results demonstrated accurate position tracking, consistent rendering of virtual dynamics, and stable operation despite the sensing, bandwidth, and actuation limitations associated with low-cost hardware. The system is capable of reproducing meaningful changes in virtual inertia, damping, and stiffness while maintaining user controllability and interaction stability. These findings demonstrate that admittance control is feasible on an ultra-low-cost haptic platform. The proposed design provides a promising foundation for affordable rehabilitation and assessment technologies and supports the development of accessible robotic tools for objective evaluation of post-stroke motor impairments. ... Read more

Accurate classification of handheld object weight during different human motion is crucial for applications in health monitoring, injury prevention and exoskeleton systems. This study investigates the feasibility of using only a single forearm mounted inertial measurement unit (IMU) combined with AI algorithms to classify both movement types and the weights of handheld objects. A series of experiments with one subject were conducted to collect IMU data under various combinations of movements and object weights. Multiple feature extraction techniques including time, frequency, and time-frequency domains were applied, followed by classification using machine learning methods (SVM, KNN) and deep learning models (1D-CNN-LSTM, Wavelet-CNN). A genetic algorithm was used for optimal feature selection in machine learning pipelines, while open set classification capability was implemented using the Convolutional Prototype Network (CPN). Result shows that deep learning models, particularly 1D-CNN-LSTM method, outperform machine learning methods, achieving up to 94% classification accuracy. Moreover, the CPN model effectively rejected unknown movement patterns in open set scenarios. The proposed framework shows promising potential for wearable systems capable of intelligent workload classification in real world environments. ... Read more

This thesis presents the development of a bidirectional, twisted-string actuated 3D-printed metamaterial hand orthosis aimed at supporting patients with impaired hand function in performing activities of daily living (ADL). Building upon a hybrid orthosis concept developed in previous work, the design was improved to address limitations such as mechanical inefficiency, spring buckling, and user discomfort, as well as actuate the device for future ADL support application. A novel actuation strategy using a one to one gear system with pre-wound wires, allowing for bidirectional motion and enhanced responsiveness. The actuation system was designed to deliver at least 120 𝑁 of linear force at 80 𝑚𝑚 of displacement, meeting the 15 𝑁 fingertip force requirement for ADL assistance. A closed-loop control system based on PID feedback was integrated, enabling the orthosis to track position targets.
In parallel, the orthosis design was refined to reduce internal friction, improve the input-output force gradient, and minimize twitching during actuation. The redesigned prototype demonstrated a 23.7% improvement in mechanical efficiency compared to the original version. Testing revealed that the system could perform a full hand closing/opening motion within 4 seconds and accurately replicate natural finger trajectories. Retesting the original orthosis under synchronized input/output conditions produced a reliable baseline
dataset, enabling a clear comparison between the legacy and updated systems. While the redesigned prototype encountered minor issues such as spring buckling under high loads due to fabrication tolerances, the overall performance met the core design goals. This work demonstrates the feasibility of integrating the novel bidirectional twisted-string actuation
system with 3D-printed metamaterial structures for wearable assistive devices. It provides a validated platform for further development toward user-ready robotic rehabilitation gloves with improved comfort, responsiveness, and usability. ... Read more

Construction and industrial workers are at high risk for work-related musculoskeletal disorders, often caused by repetitive tasks and lifting heavy loads. To reduce these work-related musculoskeletal disorders, guidelines on human strength capabilities should be followed to help mitigate injuries by reducing muscle overloading. Current solutions focus on rehabilitation and are often focused on the shoulder, elbow, and fingers, leaving the wrist vulnerable. In work-related injuries, the wrist is one of the most commonly injured body parts and is associated with high cost; both medical costs and cost due to productivity, as the median absence of work is 13 days. Injuries can be the result of bone impact forces, but more commonly muscle tissue and central nervous system damage resulting from repetitive tasks. Research shows that injuries are most common at workstations with frequent wrist deviation. Therefore, the goal of this paper is to design an orthosis to support employees in construction or industry work when working with a drill.

For the orthosis to be accepted by the users it has to be easy to use and should not limit freedom of movement. A force analysis of the wrist when holding the drill, and when using the drill against a vertical wall indicates that both the ulnar deviation and the radial deviation must be supported. To clarify which requirements are a basic necessity for the orthosis to be functional and which can be scored on functionality, they are divided into product requirements and user requirements, respectively.

Five concepts are presented to help support both the ulnar deviation and the radial deviation. Three of them are passive solutions, the other two are active. A prototype of all concepts is made and they are scored following the User Requirements utilizing a Harris Profile. During testing it is clear that the ulnar deviation support has little to no impact on the experienced muscle force. Therefore, this requirement is ignored and the orthosis ar scored with the other requirements. Both the concept using Bowden cables and the Support Arm scored well. Since the Support Arm is less complex and functional for the purpose of this research, the Support Arm is further developed.

A force analysis confirmed that the Support Arm can assist in lifting the drill by generating a configurable support force, adjustable via the spring constant or position of the attachment point. Testing the orthosis indicated that the forearm is not in line with the drill, as was expected, reducing the force on the drill. To resolve this, the attachment points of the spring and beam are angled 20 degrees to counteract the angle of the forearm compared to the drill.

The calculation in this paper are simplified, further analysis is needed to determine the effects excluded in this paper. The prototype of the orthosis should also be tested using EMG and a musculoskeletal model. This can give an indication on the effect of the orthosis on muscle activation an muscle force. More improvements of the orthosis are needed to improve user comfort and increase the force transfer. The Support Arm functions only to help lift an object. For more general wrist support and with further research, the Bowden Cable concept is promising. While more complex, the concept can help with dynamic forces to assist to the wrist movements.

In conclusion, the Support Arm orthosis effectively reduces the experienced muscle load during radial deviation when a drill is being lifted. The supporting force of the orthosis can be easily modified depending on the task. For the goal of this research, to support employees in construction and industry work when working with a drill, the Support Arm is a functional solution that can decrease the necessary muscle force and, in doing so, also decrease the bone contact forces in the wrist. ... Read more

Effective Human-Machine Interaction (HMI) depends on accurately capturing and interpreting information from the human user. For systems relying on hand-based operation, understanding the limits of human cognitive-motor abilities is crucial to designing intuitive and efficient interfaces. This study presents an experimental setup involving four analog buttons with a minimally complex control, where human hand performance is assessed through simultaneous multi-finger Fitts' Law tasks. Initially, task difficulty was assumed to be computed by summing the individual indices of difficulty for each button, which resulted in peak throughput performance with two fingers. However, this approach did not align with Fitts' Law. By applying a weighted summation, with weights based on the variation of distances within a task, the difficulty measure better conformed to Fitts' Law, and the highest throughput was achieved with a single finger. These findings highlight the interdependence in multi-finger movement complexity and emphasize the importance of considering cognitive-motor limitations when designing HMI interfaces to optimize user performance. ... Read more

Stroke survivors often exhibit motor impairments, which hinder activities of daily living. While grounded robotic perturbation devices can accurately quantify joint dynamics via system identification, their size and fixed positioning limit functional assessments under realistic conditions. To address this gap, we present the design and pilot evaluation of a novel, wearable perturbation device capable of delivering ungrounded force perturbations to the user’s forearm.

The device uses a linear solenoid actuator housed in a wrist brace to generate short, pulse-type forces, thereby inducing small angular deflections (approximately 1–3°) to the arm. An inertial measurement unit (IMU) placed on the brace tracks the resulting movement, while an accelerometer on the solenoid coil measures the perturbation force. Nine healthy participants performed three tasks—relax, resist, and move. Random pulse signals were used to prevent anticipation of the perturbations. The device successfully deflected the arm in all tasks. The largest deflections was recorded during the relax task and smaller, though still measurable, deflections in the resist and move tasks.

Estimated stiffness values in each task indicated that the device could distinguish different levels of joint rigidity, although comparisons with established literature showed some over- or underestimation. Factors such as non-rigid brace attachment and off-center actuator placement contributed to these discrepancies. Despite these limitations, the prototype demonstrates the feasibility of wearable, ungrounded force perturbations for assessing elbow dynamics. Future work will focus on improving the device’s rigidity, exploring multi-degree-of-freedom perturbations, and refining stiffness estimation algorithms to better capture realistic joint behaviors. ... Read more

Field hockey sticks are interesting hand-held sports equipment, due to their duality in preferred stick behaviour. The stick is used for striking, where a high power is desired; but also for stopping, where good control is required. This report aims to identify the properties that influence stick performance and to design a field hockey stick with adaptable properties to improve stick performance; where performance is defined as the ability of the stick to develop a high velocity when hitting a ball, and to provide proper control when stopping the ball.
The stiffness, damping and mass of the stick are properties influencing the stick behaviour; these properties are present locally, at the impact location, as well as over the full length as deflective properties due to the moment originating from the ball impact.
The deflective stiffness and damping properties are identified by applying a disturbance force on the stick tip and measuring the displacement; this shows a range of stick stiffness from 1.4 to 3.0 kN/m and a stick damping from 0.5 to 2.7 Ns/m. Measurements are performed analysing the influence of stiffness, damping, mass and effective mass; this is done by a setup where a stick falls down towards a ball and the ball distance is measured. Additionally, a mathematical model is developed for the analysis of these stick properties. This consists of a collision model, including the coefficient of restitution reflecting the stiffness and damping properties. It can be concluded that the effective stick mass is most influential and the desired properties are opposite for striking and stopping.
A design of a mechanism that fits inside the stick is proposed, this mechanism reacts to the angular acceleration of the stick, and thereby changes its properties between striking and stopping a ball. It adapts the effective mass of the stick, by two weights moving towards the head of the stick when striking a ball. By this increase in effective stick mass, an increase of 7% (compared to the original effective mass) of the ball velocity after hitting the ball is expected. ... Read more

The Shoulder Elbow Perturbator (SEP) is a robotic diagnostic device developed to assess multiple forms of motor impairment common in stroke patients. As an active medical device, the SEP would be bound to strict regulations if brought to market. By replacing its motor with a passive power source, this regulatory burden can be minimized, with the added advantage of greatly minimizing its cost. As a first step in developing this passive SEP, a prototype capable of reproducing one of the SEP’s basic tests was conceptualized. After evaluating multiple options for the passive energy source and associated components, a cost-effective design was developed using a spring, a variable radius winding drum to convert the spring’s force output to a constant torque, and a bicycle disc brake. A simulation of this design was then modeled and run through a variety of scenarios as a theoretical validation of the concept. The results of this process are promising, though testing with a physical prototype is needed for further validation. ... Read more

Background and Objective— Accurate measurement of individual finger forces is essential for assessing hand function and understanding motor impairments, yet conventional tools such as dynamometers measure only total grip strength and offer no insight into finger-specific contributions. This thesis addresses this gap by designing and validating a modular measurement system capable of measuring forces from all five digits across three predefined grasp types, operable by post-stroke patients and compatible with integration into perturbation platforms such as the Shoulder–Elbow Perturbator (SEP).

Methods and Results— An iterative Design Thinking approach guided the development of a modular hand-function assessment device featuring interchangeable grasp interfaces with spring-guided pistons to enable natural grasp motion while capturing individual finger forces. Sensor calibration using standardized weight steps showed excellent linearity, with an average calibration error of 0.15 N, substantially outperforming the 1.25 N error specified by the manufacturer. Validation with thirteen healthy participants demonstrated high repeatability across all grasp types, with SD, CV, and RMSE values comparable to a reference dynamometer. The device further reproduced expected biomechanical force-distribution patterns, including the characteristic flattening of finger contributions in larger cylindrical grasps.

Conclusion— The results demonstrate the feasibility of a practical, compact, and modular multi-finger force measurement system capable of detailed hand-function assessment. With refinement of mechanical tolerances and subsequent clinical validation, the device has strong potential for both rehabilitation assessment and integration into perturbation-based motor-control research, providing a more complete understanding of individual finger contributions during functional grasping. ... Read more

Work-related Musculoskeletal Disorders represent the largest occupational health burden in the European union, imposing significant economic and societal burdens. In this study, an ergonomic risk assessment at Bilfinger SE identified the scaffolders' passing-on task, involving repetitive vertical transfer of heavy materials, as the highest-risk activity among their services.

Additionaly, in-situ markerless kinematic (OpenCap) and kinetic (acceleration and forces) captures, as well as \textit{in-silico} biomechanical analysis (OpenSim) were combined to uncover the highest-impact areas of the passing-on task on three experienced scaffold workers.

Inverse Dynamics analysis revealed that when external forces (stander weight) were included, mean lumbar lateral bending and rotation moments increased by over 10%. Shoulder flexion moments exceeded 60 degrees regularly, with the dominant arm experiencing over 15% higher mean moments over the non-dominant arm. Smaller materials (ledger, console) resulted in average mean joint moments that are 25% higher compared to the stander: 39.8 and 40.3 Nm compared to 31.3 Nm. The mean shoulder flexion moment in the dominant arm was 55% (ledger) and 20% (console) higher compared to the non-dominant arm.

These findings highlight excessive lumbar lateral bending and rotation, and asymmetric arm usage as key risk factors for WMSDs.

This study demonstrates the feasibility of a combined in-situ andin-silico approach of uncovering high-impact areas within occupational workers, informing targeted interventions.

Future works should address OpenCap's limitations, explore the biomechanical effects of compensatory movements, and the develop interventions to mitigate the identified risks. ... Read more

Work-related musculoskeletal disorders (WRMSDs) are a leading cause of work leave in the construc tion industry. As Strukton experiences a shortage in workers for rail catenary construction, minimizing the occurance of WRMSDs is important. This is done by the implementation of new working methods or new worker aids, such as boom lifts and exoskeletons. Strukton has experienced that not all worker aids are, however, as helpful as expected. In order to find the right tools, a better understanding of the work and physical load is needed. This thesis aims to identify which parts of the catenary construction workarethemostdemandingbycombininganergonomicriskassessment(ERA)withthedevelopment of an in situ biomechanical analysis tool. The research consists of two major components. First, a qualitative ergonomic risk assessment was conducted to identify the most physically demanding tasks in rail catenary construction. Methods in cluded site visits to Strukton’s rail construction projects, unstructured and semi-structured interviews with workers and experts, and a questionnaire administered to 12 workers. These data were used to develop a Work Breakdown Structure (WBS), categorize tasks based on their physical demand, dura tion, and frequency, and identify prevalent ergonomic risks. Tasks such as wire tensioning, removing wires, and task related to the support portal were identified as high-risk activities. Back, shoulder, and ankle complaints are the most commonly reported by workers and might be related to these tasks. Fu ture research should expand the worker sample size, investigate the causes of WRMSDs further, and enhancethe reliability of task scoring. Recommendations for Strukton include adding sidewalks on site, redesigning boom lift platforms, introducing ergonomic tool bags, and using electric tools for tasks that require high force. The second component of the research involved designing and validating an in-situ biomechanical analysis tool. The tool leverages pose detection technology from widely available mobile devices and integrates it with musculoskeletal modeling software to capture and analyze joint kinematics and mo ments during construction tasks. This approach allows for task-specific quantification of physical loads experienced by workers, bridging the gap between qualitative insights and quantitative measurements. A pilot validation study demonstrated the feasibility of the tool in providing kinematic and joint moment data for biomechanical analysis, though further validation is needed to generalize its application. Poten tial uses of the tool include automating ERAs and validating conceptual worker aids. Future research should focus on deploying the tool in real-world working environments and refining the implementation of external force data in the modeling process. The findings of this thesis highlight the critical need for task-specific ergonomic interventions in the construction of rail catenaries. The combination of qualitative ERA and biomechanical analysis offers a comprehensive framework to identify risks and design targeted solutions. This work provides a foun dation for future research into ergonomic improvements in dynamic and physically intensive industries ... Read more

In Upper Motor Neuron Lesion (UMNL) following stroke, patients can experience increased joint impedance, resisting joint rotation and hindering functional movement. This heightened impedance in UMNL is driven by both exaggerated reflexes and increased intrinsic muscle activation through co-contraction, hypertonus, or synergies. The simultaneous presence of these mechanisms complicates clinical distinction, especially given their theorised interplay, where increased intrinsic activation would further heighten reflex responses. Separate quantification of this intrinsic and reflexive impedance and their interaction, can aid in further investigation of the pathophysiology of post-stroke joint impairment and its treatment.

This work presents the investigation of an Open Loop System Identification (OL-SID) protocol, to perform this separate quantification of intrinsic and reflexive impedance for the elbow joint. Perturbation experiments were performed with 16 healthy subjects, using multisine positional perturbations and measuring the elbow torque response. An impedance model consisting of both intrinsic and reflexive parameters was fit to the estimated frequency response function (FRF), relating perturbation angle to joint torque. It was assessed how background muscle activation, as well as the frequency and velocity of the perturbation signal, influenced the modelled intrinsic stiffness, intrinsic damping, and reflex velocity-gain.

For this, three different biceps muscle activation levels were requested from the participants in different trials; 0%, 10%, and 30% of Maximum Voluntary Contraction (MVC), as confirmed by online EMG measurements. Participants were requested to not actively resist perturbations, but only to comply with the requested biceps activation level. Furthermore, three rotational multisine perturbations with a max. amplitude of 2 degrees were applied; Wide Bandwidth - High Velocity, Narrow Bandwidth - Low Velocity, and Wide Bandwidth - Low Velocity. Cross-combination of biceps activation levels and perturbation signal resulted in 9 impedance quantifications per participant.

Increased biceps activation resulted in a significant increase of intrinsic stiffness, intrinsic damping, and the reflex-gain. This confirmed the expected relationship between muscle activation and intrinsic impedance, as well as the theorised relation between intrinsic activation and the reflex response. Unexpectedly, differences in used perturbation bandwidth or velocity showed no clear influence on identified reflex gain. This contradicts findings of reflex suppression during high-bandwidth force perturbations in tasks that require resisting these perturbations, as well as during high-velocity binary or unidirectional joint stretches. This discrepancy shows that joint system identification results are highly dependent on perturbation type and subject task, emphasising the need to align the experimental design with the clinical question at hand.

Despite some shortcomings regarding low coherence of the estimated FRFs, and necessary further research on perturbation signal properties and their effect on the reflex response, the results of this study are promising. The observed trends in fitted parameters with increased activation levels in line with physiological expectations, indicate the ability of this technique to validly identify reflexive and intrinsic joint impedance. This distinction is highly valuable for advancing investigation of the pathophysiology and clinical presentation of UMNL post-stroke, in the pursuit of adequate treatment for different patients.
... Read more

Work-related musculoskeletal disorders (WRMSDs) are prevalent among sonographers, particularly affecting the shoulder region due to repetitive and static movements. This study introduced an ergonomic posture and a mobile arm support (MAS) to reduce the load of muscles contributing to the stabilization of the shoulder. The experimental setup replicated a conventional cardiac ultrasound examination, using a phantom model and simulations to mimic cardiac ultrasound procedures. Professional cardiac sonographers were instructed to acquire three cardiac views (parasternal long-axis view (PLAX), apical four-chamber view (A4C), and subcostal window(SCW)) while surface EMG and joint angles of the shoulder, elbow, and back were measured. Additionally, subjects were tasked with completing a questionnaire to gather subjective outcomes of usability and satisfaction with the support and ergonomic posture. Muscular activity of the middle deltoid muscular activity (PLAX; F(3,12)10.15,p=.001,η2 p=.717, A4C; F(3,12)=5.75,p=.011,η2 p=.590) and superior trapezius (PLAX; F(3,12)=7.05,p=.005,η2 p=.638) decreased significantly during examinations with postural changes but increased significantly while only support was provided. The support was considered slightly useful (SUS=65.0±6.9, α=.585), but the ergonomic change was considered poor (SUS=43.0.0±14.4, α=.813). The increase in muscular activity was likely caused by incorrect placement of the MAS on the upper extremity and incorrect levels of support. Besides, sonographers reported a MAS would be useful when there is an additional functionality that applies contact force through the MAS. The addition of support to postural changes did not significantly reduce muscular load compared to examinations with only postural changes. Therefore, ergonomic postural change is a sufficient solution for mitigating WRMSDs development due to the only significant decrease in muscular activity. ... Read more

Human intention detection with hand motion prediction is critical to drive the upper-extremity assistive robots. However, the traditional methods relying on physiological signal measurement are restrictive and often lack environmental context. We propose a novel approach that integrates gaze information, historical hand motion sequences, and environmental object data to predict future sequences of intended hand poses, adapting dynamically to the assistive needs of the patient without prior knowledge of the intended object for grasping. Specifically, we propose to use a vector-quantized variational autoencoder for robust hand pose encoding with an autoregressive generative transformer for effective hand motion sequence prediction. We demonstrate the usability of these novel techniques in a pilot study with healthy subjects. To train and evaluate the proposed method, we collect a dataset consisting of various types of grasp actions on different objects from multiple subjects. Through extensive experiments, we demonstrate that the proposed method can successfully predict sequential hand movement. Especially, the gaze information shows significant enhancements in prediction capabilities, particularly with fewer input frames, highlighting the potential of the proposed method for real-world applications. ... Read more

This research investigates methodologies to reduce and implement negative inertia in robots for upper extremity diagnostics and rehabilitation. The robot’s responsiveness is enhanced by integrating accelerometers and Kalman filters into the control scheme, ensuring smoother physical human-robot interactions. A force gain that mimics negative inertia significantly improves system dynamics within the admittance control framework. Introducing dead zones for force and acceleration stabilizes responses at lower rendered inertia, crucial for handling spastic conditions. However, this research identifies a lack of standardized evaluation methods for negative inertia and highlights hardware constraints, such as bandwidth limitations, that restrict performance. Future research should focus on establishing evaluation standards and optimizing hardware to refine control precision. This work demonstrates the potential of advanced control strategies to optimize robotic rehabilitation, paving the way for more effective diagnostic and therapeutic interventions. ... Read more

This paper aims to explore innovative fabrication methods for assistive gloves using origami design principles. The study addresses the limitations of current assistive gloves, such as bulkiness and lack of customisation, by developing a novel fabrication method that improves ergonomic design and manufacturing flexibility. The proposed method integrates 3D printing and welding techniques to create lightweight, customisable gloves that enhance rehabilitation for individuals with hand impairments, particularly stroke survivors. The paper details the design process, and testing methods, concluding with successfully creating a new fabrication approach for origami-based assistive gloves. ... Read more

This thesis explores the application of origami-inspired inflatable structured sheets to enhance muscle function for individuals with muscle impairments, specifically focusing on the biceps brachii muscle. These sheets induce muscle contraction through pneumatic pressure, aimed at facilitating forearm flexion by strategically incorporating patterns that create air pockets. These sheets hold significant potential for assisting individuals with conditions such as muscular dystrophy, cerebral palsy, or spinal cord injuries. Compared to traditional solutions, they offer less bulkiness and greater adaptability, promoting natural movement and improving limb mobility. Furthermore, these sheets can be customized to fit various muscle contours and integrated with mechatronics for real-time adjustments, enhancing stability and resilience in muscle support applications. The study begins with the development of a comprehensive test setup that simulates muscle extension and flexion, integrating mechanical and electrical components. This setup includes a silicone muscle model of the biceps brachii muscle, essential for evaluating sheet performance. Subsequent chapters delve into the evaluation of materials and patterns for the sheets. Materials such as PVC-film, TPU-film, PET-film, and Nylon are assessed for biocompatibility, flexibility, temperature sensitivity, and adhesion characteristics. Here, TPU-film emerges as the most promising material due to its durability and adhesive properties under pressure. For the evaluation of the patterns, a preliminary study is conducted to determine optimal patterns that effectively respond to pneumatic pressure for inducing muscle contraction. This study reveals that combinations of parallel lines and zigzag structures show the most promising results among the tested patterns. Building on these findings, various structured sheets are tested around the muscle model to assess their ability to activate and contract muscles. An optimal structure is selected that conforms closely to the contours of the muscle model, ensuring optimal airflow and maximizing the potential for effective muscle contraction. However, challenges such as leakage affecting pressure retention and muscle activation are identified, underscoring the need for further optimization to achieve practical muscle activation and flexion. In conclusion, while this research provides valuable insights into the feasibility of origami-inspired structured sheets for muscle augmentation, ongoing refinement is crucial to address practical challenges and optimize performance. ... Read more

Fear of pain is a psychological consequence that can hinder an athlete’s ability to resume their previous level of performance or even return to their sport. A desire exists to examine the influence of fear of pain on both mechanical and neural factors involved in postural control, particularly focusing on the lower extremities. Previous literature shows associations between jump tests and fear of pain, but existing methods cannot differentiate between neural and mechanical factors, necessitating a different approach. Given the absence of a method to measure lower extremity reflexes and intrinsic properties similar to those observed in the shoulder, the original Proprio setup was adapted for leg perturbation to investigate its impact on lower extremity postural control. The adapted setup consists of a platform for participants to stand on, with coupling provided by a stiff ankle brace. A group of seven participants performed a slack trial and multiple stiff trials in both angled and neutral positions. Metrics such as admittance, stiffness, and coherence were calculated for each trial. The suitability of the setup was evaluated based on the selectivity of results across different trials and positions. Additionally, coherence was assessed, indicating the linearity which is necessary for future parameter identification. Notable differences in admittance and stiffness were observed between different legs and trials. This demonstrates the ability of the new setup around the Proprio to differentiate between various conditions. Comparisons with previous studies on shoulder dynamics showed similar findings, suggesting accurate measurement of leg dynamics. High coherence values indicated linearity, supporting the potential for future estimation of intrinsic parameters and reflex gains in leg dynamics. Considering these factors, the adapted Proprio setup appears suitable for analyzing lower extremity postural control. The next steps include expanding the participant group for significance and parameter identification, which will enable the investigation of the impact of fear of pain on intrinsic and reflexive properties using this setup. ... Read more