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

Sleep is crucial in rehabilitation processes, promoting neural plasticity and immune functions. Nocturnal body postures can indicate sleep quality and frequent repositioning is required to prevent bedsores for bedridden patients after a stroke or spinal cord injury. Polysomnography (PSG) is considered the gold standard for sleep assessment. Unobtrusive methods for classifying sleep body postures have been presented with similar accuracy to PSG, but most evaluations have been done in research lab environments. To investigate the challenges in the usability of a previously validated device in a clinical setting, we recorded the sleep posture of 17 patients with a sensorized mattress. Ground-truth labels were collected automatically from a PSG device. In addition, we manually labeled the body postures using video data. This allowed us also to evaluate the quality of the PSG labels. We trained neural networks based on the VGG-3 architecture to classify lying postures and used a self-label correction method to account for noisy labels in the training data. The models trained with the video labels achieved a higher classification accuracy than those trained with the PSG labels (0.79 vs. 0.68). The self-label correction could further increase the models' scores based on video and PSG labels to 0.80 and 0.70, respectively. Unobtrusive sensors validated in clinics can, therefore, potentially improve the quality of care for bedridden patients and advance the field of rehabilitation.

Brightness-mode (B-mode) ultrasound has been used to measure in vivo muscle dynamics for assistive devices. Estimation of fascicle length from B-mode images has now transitioned from time-consuming manual processes to automatic methods, but these methods fail to reach pixel-wise accuracy across extended locomotion. In this work, we aim to address this challenge by combining a U-net architecture with proven segmentation abilities with an LSTM component that takes advantage of temporal information to improve validation accuracy in the prediction of fascicle lengths. Using 64,849 ultrasound frames of the medial gastrocnemius, we semi-manually generated ground-truth for training the proposed U-net-LSTM. Compared with a traditional U-net and a CNNLSTM configuration, the validation accuracy, mean square error (MSE), and mean absolute error (MAE) of the proposed U-net-LSTM show better performance (91.4%, MSE =0.1± 0.03 mm, MAE =0.2± 0.05 mm). The proposed framework could be used for real-time, closed-loop wearable control during real-world locomotion.

Continuous monitoring of muscle coordination can provide valuable information regarding an individual's performance during physical activities. For example, changes in muscle coordination can indicate muscle fatigue during exhaustive exercise or can be used to track the rehabilitation progress of patients post-injury. Traditional methods to evaluate coordination often focus solely on measuring muscle activation with electromyography, ignoring timing changes of the resultant force produced by the activated muscle. Setups designed to evaluate force directly to study muscle coordination are often limited by either hyper-constrained settings or cost-prohibitive hardware. In this paper, we employ wearable, ultra-sensitive soft strain sensors that track muscle deformation for estimating changes in muscle coordination during cycling at different cadences and to exhaustion. The results were compared to muscle activation timing measured by electromyography and peak force timing measured by a cycle ergometer. We demonstrate that with an increase in cadence, the soft strain sensor and ergometer timing metrics align more closely than those measured by electromyography. We also demonstrate how muscle coordination is altered with the onset of fatigue during cycling to exhaustion.

The loss of sensitivity of the upper limb due to central or peripheral neurological injuries severely limits the ability to manipulate objects, hindering personal independence. Non-invasive augmented sensory feedback techniques are used to promote neural plasticity hence to restore the grasping function. We devised a wearable device for hand sensorimotor rehabilitation capable of reliably detect transient tactile events based on custom piezoelectric polyvinylidene fluoride (PVDF) sensors and deliver discrete bursts of vibrations upon these events. We integrated the sensors into a fabric glove and tested the device in a pilot bench test exploring its ability to detect object contact and release as well as object slippage. Due to their broad bandwidth, the sensors proved to be suitable for both the applications: they responded with clear peaks when touching or releasing the object and increased the high-frequency content of the signal during slippage.

For those with upper limb absence, body-powered prostheses continue to be popular for many activities despite being an old technology; these devices can provide both inherent haptic feedback and mechanical robustness. Yet, they can also result in strain and fatigue. Body-powered prosthetic graspers typically consist of a simple lever providing a relatively constant transmission ratio between the input forces from the user's shoulder harness and the grip force of their prosthetic prehensor. In the field of robotic hand design, new continuously varying transmissions demonstrate particular promise in generating a wide range of grasping speeds without sacrificing grip strength. These benefits, if applied to shoulder-driven prosthetic grippers, have the potential to both reduce shoulder exertion and fatigue. This work presents the integration of a continuously variable transmission into a body-powered, voluntary close prosthetic testbed. We introduce the design and validate its performance in a benchtop experiment. We compare constant transmission conditions with a force-dependent, continually varying condition. The device is mounted on a prosthetic emulator for a preliminary wearable demonstration.

Powered lower limb exoskeletons have been used in recent decades to support and improve conventional gait rehabilitation programs. In this context, visual feedback is considered a valuable tool to facilitate patient learning and engagement. Treadmill-based lower limb robotic exoskeletons are commonly incorporated with traditional screens or virtual reality (VR) devices. However, these modalities are incompatible with over-ground robotic exoskeletons, in which users should pay attention to multiple elements of the open environment and more complex tasks. Recent advances in augmented reality (AR) technology have unlocked a new way of displaying 3D graphics in untethered wearable devices like Microsoft's Hololens 2 without occluding the rest of the user's field of view. These advances can be crucial in certain situations and position AR as an excellent candidate to improve visual feedback when using robotic exoskeletons. In this work, we present the development of an AR-based audio-visual feedback system that tracks the trunk rotation from an Inertial Measurement Unit (IMU) to aid in walking tasks assisted by a lower limb robotic exoskeleton. A preliminary study was done integrating the H3 robotic exoskeleton with Hololens 2. We observed that displaying posture information is feasible and could help mitigate the unnatural posture often imposed by a robotic exoskeleton.

Young children with motor disabilities face extra obstacles to engaging in movement and initiating social interaction. A body-weight support harness system (BWSS) allows a child to take steps, explore the environment, and interact with people and objects, but further research is needed to understand how this type of system can help children with motor disabilities. Assistive robots have the potential to keep a child engaged and motivated during physical therapy sessions with a BWSS. We conducted a case study over three and a half months to understand if the BWSS alone and if the BWSS with an assistive robot could promote child movement and engagement. Our results show that the child tended to increase their amount of movement over each session with the BWSS. The assistive robots used in this study also tended to keep the child engaged. The products of this work can benefit clinicians and researchers interested in early mobility intervention technologies, as well as roboticists working in the child mobility domain.

Sleep is essential to boost the rehabilitation outcome as it facilitates motor learning, enhances cognitive performance, and improves mood and well-being. Rocking beds that provide vestibular stimulation may be a promising and non-invasive alternative to conventional pharmaceutical treatments for individuals with sleep problems, offering regenerative sleep without unwanted side effects. Previous research has shown that the effectiveness of the interventions is related to the chosen rocking acceleration. Moreover, the movement of the bed must be comfortable and smooth to avoid disturbing the user's sleep. Previously, the motor control parameters were tuned manually, which was time-consuming, subjective, and did not guarantee minimum deviation from the desired acceleration profile. In this work, we present an efficient and effective method using Gaussian processes to automatically tune the PI control parameters of a rocking bed moving along the longitudinal axis. We first simulated the kinematics of a rocking bed and optimized the control parameters for a chosen objective function that included the desired and the actual accelerations in the movement direction. We then compared the number of iterations needed to reach this objective for a model based on Gaussian processes and for a model based on a naive random exploration of the parameter space. Finally, we implemented the Gaussian process on the rocking bed to automatically tune the control parameters and subjectively compared them to the control parameters that were previously obtained after manual tuning. Our simulation showed that we can reach the control objective after a constant number of iterations using Gaussian processes, independent of the search space size. For the random search, the number of iterations increased quadratically with the size of the search space. The Gaussian process was found to be well transferable to the rocking bed. After less than one hour, control parameters were discovered that outperformed the previous parameters in terms of smoothness. However, despite the smoother motion, the noise emission from the motor, which was not part of the optimization, increased considerably. Our presented technique based on Gaussian processes significantly reduced the time and effort required to optimize the bed's control parameters compared to manual tuning. In future work, the control objective has to be refined to include noise emission as an optimization metric as low noise is an important aspect in sleep-related applications.

Independent physiotherapy at home is a crucial element of rehabilitative care for a wide range of conditions as it constitutes a large portion of the overall therapy dose. However, up to 80% of individuals who are prescribed at-home physiotherapy do not consistently adhere to their treatment schedule, resulting in poor treatment outcomes. This is likely due to a lack of motivation and progress tracking in the current standard of care. We have developed a novel software prototype that allows users to control commercial entertainment content, such as video games or interactive music videos, with their movements during physiotherapy. By connecting therapy to proven entertainment content, we aim to improve on the current motivational deficits. This study investigated the safety and feasibility of this concept in a controlled environment over four physical therapy sessions with seven patients suffering from musculoskeletal and neurological conditions. As a secondary outcome, patients were asked about their enjoyment, perceived competence and effort using the Intrinsic Motivation Inventory (IMI) questionnaire. All participants were able to interact with the presented entertainment content and completed the study with no adverse events. Despite the diversity in pathology, age and training scenarios, the entertainment content maintained the patients' enjoyment with a high average rate of 6/7 on the IMI scale. Interacting with commercial entertainment content by doing physical therapy exercises was feasible, safe, and well-received over the six-week study period.

Gait impairments is a common condition in post-stroke subjects. We recently presented a wearable ankle exoskeleton called G-Exos, which showed that the device assisted in the ankle's dorsiflexion and inversion/reversion movements. The aim of the current pilot study was to explore spatiotemporal gait parameters and center of pressure trajectories associated with the use of the G-Exos in stroke participants. Three post-stroke subjects (52-63 years, 2 female/1 male) walked 160-meter using the G-Exos on the affected limb, on a protocol divided into 4 blocks of 40-meters: (I) without the exoskeleton, (II) with systems hybrid system, (III) active only and (IV) passive only. The results showed that the use of the exoskeleton improved swing and stance phases on both limbs, reduced stride width on the paretic limb, increased stance COP distances, and made single support COP distances more similar between the paretic and non-paretic limb. This suggests that all G-Exos systems contributed to improving body weight bearing on the paretic limb and symmetry in the gait cycle.