IEEE ... International Conference on Rehabilitation Robotics : [proceedings]最新文献

Shoulder kinematics has been required to be estimated using compact sensors to provide visual feedback on joint movements for patients undergoing home rehabilitation due to musculoskeletal diseases, such as stroke. Multiple strain patches have been used for shoulder kinematics estimation, either embedded in tight shirts or attached directly to the skin. However, it is desirable to eliminate the tight shirt while minimizing the number of strain patches for convenient use in home rehabilitation. This study presents a single all-in-one skin-adhesive strain patch that enables shoulder kinematics estimation through placement optimization. At the optimal on-skin position, strains caused by flexion-extension and abduction-adduction movements were generated significantly in different directions. The fabricated all-in-one strain patch measured these strains using embedded sensors oriented accordingly. Moreover, the patch was designed to be lightweight and ultra-thin, allowing for convenient attachment to the skin. Shoulder kinematics estimation using this strain patch was preliminarily evaluated with a subject, demonstrating that shoulder joint angles could be estimated with an average error of 15.8 % of the range of motion for two degrees of freedom. The placement optimization reduced the total number of strain patches required for kinematics estimation, enhancing convenience for daily use.

Reactive balance responses, which involve corrective and protective strategies, are highly dependent on rapid muscle activation to restore postural stability. Although electromyography (EMG) is commonly used to measure muscle activity, it has limitations such as signal interference, particularly during fast responses to external disturbances. Ultrasound imaging (US), in contrast, provides visualization of both superficial and deep muscles. Combining EMG and US imaging offers complementary insight into muscle behavior during reactive balance tasks. In this study, we investigated muscle activation and fascicle length changes in the medial gastrocnemius (MGS), lateral gastrocnemius (LGS), and soleus (SOL) muscles of the dominant (stepping) leg during stepping responses to unexpected low-amplitude ($57.6 pm 5.8 ~mathrm{N}$) and high-amplitude ($123.4 pm 11.1 ~mathrm{N}$) waist-pull perturbations in the anterior and posterior directions. Five young male adults (age: $25.2 pm 5.5$ years) participated in the study. Results showed that perturbation amplitude significantly affected the EMG activation of both the MGS and SOL muscles in both directions, consistent with previous studies. Similarly, perturbation amplitude impacted fascicle length shortening in the LGS and SOL muscles. Significant differences in MGS and SOL activation were observed between high-amplitude and lowamplitude perturbations in both directions. Fascicle shortening in the LGS also differed significantly between perturbation amplitudes, whereas SOL fascicle shortening did not. By combining EMG and US imaging within the same participants, this study provides new insights into the neuromuscular mechanisms underlying balance control. These findings may inform the development of improved control strategies for neurorehabilitation devices and fall-prevention systems.

The human hand is an exceptionally complex anatomical structure with 31 degrees of freedom. In the development of dynamic wrist-hand orthoses, the wrist's multiaxial mobility is of particular interest. To accurately replicate these natural movements, a promising approach is the use of compliant tensegrity structures. Tensegrity structures allow pivot points and axes of rotation to align with anatomical positions, enabling unrestricted mobility in all directions without conventional joints. Additionally, they permit customized movement restrictions based on therapeutic needs. The orthosis's minimalist, lightweight design ensures both effective joint stabilization and free access to key injured regions. Customizing each orthosis to the patient's unique anatomy and functional needs is crucial to prevent unnecessary strain from improper positioning. The positioning of the orthosis is directly linked to the forces applied to the wrist while using. A precise understanding of the behavior of the orthosis and its influence on the wrist forces present is therefore essential. This work explores key aspects of tensegrity-based orthosis development, emphasizing accurate 3D scanning of hand anatomy, initial experimental measurements, and simulated calculations. The proposed methodology provides a solid foundation for the design of initial prototypes of tensegrity-based hand orthoses.

Unsupervised robot-assisted therapy could allow increasing upper limb therapy dose for stroke survivors with minimal additional burden on the healthcare system. Thanks to the ability to actively assist movement and dynamically adapt the assistance level, actuated devices can support individuals with a wide range of deficits. However, these devices are often complex to use, and their application in a fully unsupervised setting has rarely been explored. Here, we present a pilot study investigating the feasibility of unsupervised therapy with ReHandyBot, an actuated device for upper limb rehabilitation. The increase in therapy dose achieved during unsupervised training, device usability and user experience were evaluated. Stroke inpatients of a rehabilitation clinic learned how to use the device for two weeks with progressively decreasing levels of supervision. After discharge, they could take it home for two weeks of unsupervised therapy. Four of the five recruited participants learned how to use the device without supervision, and three completed the protocol. During the two weeks at home, on average they performed 518.3 minutes of therapy with ReHandyBot. Usability and user experience ratings show that the device was well accepted. These positive results support larger studies investigating unsupervised home therapy with ReHandyBot and suggest that active devices can be used by patients with no to mild cognitive impairments at home without the supervision of external persons.

The coordination between muscles during the execution of a bilateral task can be analyzed by recurring to functional muscle networks, which indicates how human muscles are mutually synchronized - i.e., share an oscillatory common input - during the motor task. So far, no studies have studied the impact of an assistive exoskeleton on the neural substrates of human muscles. This work addresses this question by extracting functional muscle networks of healthy individuals during the execution of a bilateral dynamic task without and with the support of a shoulder exoskeleton. The outcomes have revealed a global stationarity of these networks, together with a slight change of synergies' shape, reflecting how the employed assistive device can change physiological patterns from both temporal and frequency domains. Despite the limited sample size, the authors claim the potential of functional connectivity analysis as an additional tool for the evaluation of inter-muscular coordination for such clinical purposes as rehabilitation robotics.

Soft assistive wearable robotics, or soft exosuits, have shown great potential in enhancing human motor function while preserving the user's natural movement. They are increasingly used to reduce fatigue and the risk of injury during physically demanding tasks and everyday activities. A deeper understanding of how assistive forces from soft exoskeletons influence muscle coordination during various tasks is crucial for developing optimized control and rehabilitation strategies for these devices. This study investigates muscle synergies in healthy participants carrying loads, as often performed in daily activities, over an extended period, both with and without the assistance of a soft elbow exosuit. Synergies similarity was analyzed by comparing the extracted patterns in both conditions, as well as their composition in terms of muscle contributions. The results show that, on average, in the with-exosuit condition fewer synergies are required to explain the performed movements. This suggests the idea that the assistive device "aggregates" motor primitives while reducing muscle activation. Furthermore, we found that synergies dominated by the wrist flexor, biceps, and infraspinatus were not altered by the exosuit assistance. However, the deltoid-dominated synergy observed during load carrying without the exosuit was altered by the exosuit assistance, shifting to a triceps-dominated synergy.

Movement smoothness is a commonly employed metric of motor impairment in stroke patients due to a strong correlation with Fugl-Meyer Assessment score, and because it can be assessed with functional-oriented tasks like wrist pointing. Robotic assessments of wrist-pointing tasks typically use targets located at fixed points in the workspace, which can exceed the range of motion (ROM) of stroke patients who exhibit severe motor impairment. We hypothesize that movement smoothness assessed based on movements that do not reach visual targets may conflate assessment of movement smoothness with ROM. We propose that scaling the locations of wrist pointing targets for wrist flexion-extension and radial-ulnar movements from a ROM assessment will better reflect movement characteristics. In this paper, we analyze wrist-pointing movements for fixed and ROM-scaled target locations. We evaluate wrist-pointing performance in both conditions for 8 neurologically-intact participants whose wrist ROM was constrained with a wrist brace simulating ROM deficits present after stroke. Results show that failing to reach targets during wrist pointing has a significant negative impact on movement smoothness. Additionally, scaling target placement to ROM increases the proportion of targets reached while producing practically equivalent movement smoothness to successfully reached fixed target locations. These findings support incorporation of ROM-scaled target placement into movement smoothness assessment.

Despite the growing demand for healthcare services due to an aging population, patients may avoid traditional rehabilitation centers due to high costs, discomfort, and time constraints associated with in-person assessments. Home-based rehabilitation offers a promising alternative, but effective kinematic monitoring and assessment remain challenging, especially for real-time applications. To address this gap, we have developed real-time, full-body kinematic analysis and visualization based on 12 wearable inertial measurement units (IMUs). Full body real-time kinematics estimation (20 Hz) was evaluated during walking, running, squatting, boxing, yoga, dance, badminton, and various seated extremity exercises and IMU estimations were compared with optical motion capture to determine accuracy. Results showed that walking was the most accurate with 5.4 deg median RMSE, and the overall median RMSE was 7.2 deg for all activities. A mean of 1.0 deg RMSE against offline computations (100 Hz) was also demonstrated, with a mean latency of 44.1 ms from sensor data acquisition to kinematic output. This approach holds the potential to revolutionize rehabilitation by enabling rapid assessment and real-time biofeedback for motion performance in orthopedic and neurological conditions and could significantly enhance treatment outcomes and patient compliance.

Flexibility training involving stretching aims to increase the range of motion (ROM) about a joint and mitigate factors that limit mobility. Stretching has been used for improving human performance and preventing injuries. Further, stretching can be used as spasticity treatment in people with spinal cord injury to provide relief from muscle spasms and improve passive ROM. Therapists apply manual stretching to participants; however, providing personalized and repeatable stretching forces is caregiver intensive. This paper describes the methods to apply three stretching techniques and determines the feasibility of implementing their protocols using a powered device and open-loop Functional Electrical Stimulation (FES) patterns. This paper develops a robust kinematic closed-loop controller to emulate the static, activeisolated (AIS), and proprioceptive neuromuscular facilitation (PNF) stretching techniques targeting the hamstring complex in which the human lies in a supine position. The robust electric motor controller rotates the leg from its initial position through a desired ROM using a rigid pivot arm actuated by a Bowden cable. When the leg reaches the target end ROM, FES inputs are applied for AIS and PNF stretching. Experiments were safely conducted in two able-bodied individuals to demonstrate the device's feasibility of implementing each stretching technique on the hamstring complex. A Lyapunov stability analysis ensures exponential tracking of the motor controller.

Physical rehabilitation exercises suggested by healthcare professionals can help recovery from various musculoskeletal disorders and prevent re-injury. However, patients' engagement tends to decrease over time without direct supervision, which is why there is a need for an automated monitoring system. In recent years, there has been great progress in quality assessment of physical rehabilitation exercises. Most of them only provide a binary classification if the performance is correct or incorrect, and a few provide a continuous score. This information is not sufficient for patients to improve their performance. In this work, we propose an algorithm for error classification of rehabilitation exercises, thus making the first step toward more detailed feedback to patients. We focus on skeleton-based exercise assessment, which utilizes human pose estimation to evaluate motion. Inspired by recent algorithms for quality assessment during rehabilitation exercises, we propose a Transformer-based model for the described classification. Our model is inspired by the HyperFormer method for human action recognition, and adapted to our problem and dataset. The evaluation is done on the KERAAL dataset, as it is the only medical dataset with clear error labels for the exercises, and our model significantly surpasses state-of-the-art methods. Furthermore, we bridge the gap towards better feedback to the patients by presenting a way to calculate the importance of joints for each exercise.