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

Micro-controlled lower limb prosthetic devices typically employ onboard sensors on the amputated side of the user's body to generate control patterns. However, balanced ambulation in a healthy subject requires synchronized neural control to facilitate appropriate positioning and orientation of both limbs. Building upon this idea, this study introduces a control architecture that utilizes information acquired from both limbs using only two inertial measurement units (one placed on the left shank, and one placed on the right thigh) to generate the control commands for a transtibial prosthesis prototype clamped to a table-top. Multiple trials were performed with a healthy subject walking on level ground at different speeds and undertaking an obstacle avoidance task. The results of the experiments suggest that incorporating data from each side provides comprehensive information about the limb's positions and orientations, thereby enabling a controller that can handle actuation timing and assist in precise prosthesis control through an understanding of the user's gait states.

This paper presents a flexible software with a Graphical User Interface (GUI) designed to assist clinicians in upper limb robotic rehabilitation sessions. The flexibility comes from the fact that, since it relies on the Robot Operating System 2 (ROS2) framework, it is potentially compatible with any collaborative industrial robot, with relatively low integration effort. The software collects and analyzes data related to each session to identify critical movement directions for each patient, detecting improvements or regressions through statistical tests, thus reducing subjectivity. Moreover, it suggests personalized physical and cognitive rehabilitation exercises focusing on the most critical movement directions. Results are visualized through intuitive graphs, allowing clinicians to design custom trajectories tailored to each patient's needs. The software has been tested in combination with a Franka Emika Panda robot proving its seamless integration and ability to process data in real time.

After a neurological adverse event, restoring trunk functionality is essential for regaining balance, proprioception, and strength, all of which are critical for performing activities of daily living and enabling rehabilitation for impaired upper and lower limbs. Traditional therapy focuses on enhancing core strength, stability, and proprioceptive capabilities. Recent advancements in the field propose robotic solutions that utilize platforms to induce specific movements, stimulating muscular activation and prompting patients to respond to guided motions. These solutions often employ robotic seats that mimic traditional clinical tools, such as cushions or physiotherapy balls, to assist in mobilizing patients. In this work, we introduce a novel approach to trunk rehabilitation through a robotic system: a Series Elastic Actuator (SEA)-based trunk robotic mobilizer. This device is paired with a suite of control strategies designed to address various neurological conditions, including spinal cord injury, post-stroke impairments, and pusher syndrome.

Robotic rehabilitation can deliver high-dose gait therapy and improve motor function after a stroke. However, for many devices, high costs and lengthy setup times limit clinical adoption. Thus, we designed, built, and evaluated the Passive Mechanical Add-on for Treadmill Exercise (P-MATE), a low-cost passive end-effector add-on for treadmills that couples the movement of the paretic and non-paretic legs via a reciprocating system of elastic cables and pulleys. Two human-device mechanical interfaces were designed to attach the elastic cables to the user. The P-MATE and two interface prototypes were tested with a physical therapist and eight unimpaired participants. Biomechanical data, including kinematics and interaction forces, were collected alongside standardized questionnaires to assess usability and user experience. Both interfaces were quick and easy to attach, though user experience differed, highlighting the need for personalization. We also identified areas for future improvement, including pretension adjustments, tendon derailing prevention, and understanding long-term impacts on user gait. Our preliminary findings underline the potential of the P-MATE to provide effective, accessible, and sustainable stroke gait rehabilitation.

Amputations of the upper and lower limbs can significantly impact psychological well-being and functional independence, limiting a person's ability to perform daily tasks. Artificial limbs aim to restore lost functionality by replicating the natural limb as closely as possible. To this aim, the ideal human-machine interface (HMI) should provide seamless connection with the natural sources of control and sensory feedback. Recent advances in surgical techniques and prosthetic technologies have enhanced residual limb functionality and enabled direct connections with peripheral neural pathways. This study introduces a simulation platform integrating the agonist-antagonist myoneural interface (AMI) with a myokinetic interface based on implantable magnets, to generate synthetic data on muscle deformation and magnet displacement as potential control signals for an assistive device. A transtibial amputation was simulated, replicating the natural agonist-antagonist mechanical connection between paired muscles. The model simulated muscle contractions and magnet implantation, which displacement was tracked using a localization algorithm. Results demonstrated a combined contraction and stretching of the agonist and antagonist muscles and the feasibility of accurately tracking implanted magnets via external sensors. These findings suggest that combining the AMI with magnet implantation could contribute to more intuitive prosthetic control in the future. In this view, this simulation platform provides a valuable pre-operative planning tool to optimize magnet placement and enhance HMI performance for individual patients.

Ankle push-off power generated by the plantar flexor muscles is essential for efficient walking but is often impaired in individuals with cerebral palsy (CP), leading to inefficient walking. Both powered and passive (i.e., motor- and spring-based) wearable resistance devices are being developed for targeted training. This study explored the use of ankle power biofeedback across powered and passive wearable resistive devices to improve muscle recruitment and push-off power in individuals with CP. Seven individuals with CP completed walking sessions under the following conditions: (1) baseline (no device), (2) spring resistance with and without biofeedback, and (3) motor resistance with and without biofeedback. Push-off power and muscle recruitment were compared between biofeedback vs no-biofeedback conditions for both devices, and to baseline. Combined spring resistance and biofeedback increased soleus activity by 40 % compared to the spring resistance only ($p=0.004$) and by 48 % compared to baseline ($p =0.002$). Similarly, spring resistance and biofeedback increased peak ankle power by 32% relative to the spring resistance only ($mathbf{p}=0.009$) and by 33% compared to baseline ($mathbf{p}=0.010$). In contrast, motorized resistance and biofeedback did not significantly increase peak soleus activity or peak ankle power relative to motor resistance only ($mathrm{p}=0.544; mathrm{p}=0.544$). These findings show that ankle power biofeedback can augment spring resistance to elicit increased muscle recruitment and power during push-off in CP.

Light Touch (LT) has been known to improve standing balance without mechanical support by providing sensory information about the movement of the body. Inspired by this, this work developed a Virtual Cane (VC) which gives no physical support but provides other sensory information that a physical cane would. The VC is developed with a distance sensor and vibration actuators to provide cane tip-to-ground distance information to the user. The extent to which this haptic feedback improves standing balance was assessed in a human experiment. 10 healthy young participants underwent a standing balance experiment with tandem stance and eyes closed, using VC with No Feedback (NF) as the baseline, VC with feedback, and with physical cane (PC). Center-of-Pressure (CoP) metrics as well as Sway-density metrics were analyzed to study the effect of these conditions on standing balance. It was found that the CoP metrics in VC were significantly improved compared to baseline (NF) and approaching the benefit of a full physical cane (PC). Sway-density metrics showed no difference between the conditions. This shows that simple, binary feedback from VC on the position of the body and cane was sufficient to positively affect standing balance without significantly altering the biomechanical strategy of standing balance. This is also a demonstration of an assistive device with a complete absence of mechanical support that can provide substantial benefit to balance.

Robotic exoskeletons offer the potential to increase gait training dosage by providing therapy outside the clinic. Exoskeleton use in naturalistic settings has been reported but under researcher supervision and for the purposes of assisting mobility. Here, we examined the feasibility of delivering a resistance-based exoskeleton gait training paradigm in the community setting. Five children aged 6-15 years with crouch gait/knee extension deficiency completed 12-weeks of gait training (5 days/week) at home with the exoskeleton as part of an ongoing interventional trial. Weekly telehealth check-ins were completed. Feasibility was assessed using patient-reported adherence to the prescribed dosage, number of adverse events, and exoskeleton hardware and software robustness during the intervention period. User satisfaction was assessed using the QUEST 2.0. No serious adverse events were reported. Participants completed between 44-57 of the targeted 60 sessions of gait training. Exoskeleton hardware and software were reliable with the most common controller adjustments involving tuning of underfoot sensing thresholds to detect stance and swing. Participant satisfaction with the exoskeleton (mean: 3.65/5) and service (4.7/5) were high. Thus, our findings indicate that resistive gait training with an exoskeleton in the real-world is safe and tolerable in children with neuromotor disorders and should be further investigated to determine effectiveness for improving functional mobility.

The integration of rehabilitation robotics and surface electromyography (sEMG) offers a powerful approach for monitoring and enhancing recovery in patients with neuromuscular disorders. However, variability in baseline sEMG readings across individuals can limit its effectiveness. Factors such as age, height, and weight influence these baselines, and there is a lack of personalized baselines that account for demographic differences. This study proposes a novel model to estimate individualized baselines for one important sEMG parameter, Root Mean Square (RMS). Demographics and physiological data were collected from 30 healthy participants, and sEMG signals were recorded using four electrodes on the forearm muscles during a pushing task at two wrist positions. A Decision Tree Regression model was developed for each combination of the two wrist postures and four electrode locations, resulting in eight combinations, with optimal features identified using the Recursive Feature Elimination method. The regression models achieved accuracies ranging from 88.81% to 95.6%. A global sensitivity analysis using the Sobol method evaluated the importance of each input feature. Results indicate that gathering more comprehensive sEMG data for the most influential factors could improve model generalizability. The findings of this study offer a promising approach for individualized sEMG baselines, with potential applications in rehabilitation robotics to enable personalized recovery strategies for neuromuscular disorders.

To secure the amount of rehabilitation in motor function recovery training after stroke, it is important to realize a rehabilitation robot for the upper limb to guide the reaching motion of patients in the acute and convalescent phases of rehabilitation in hospitals and in the maintenance phase of rehabilitation at home without a therapist. Therefore, in this study, the escorttype rehabilitation robot, which has been proposed as a structure to guide the user's hand position and arm posture, is shown to have a motion range for the user's reaching motion including arm posture. In addition, it is shown that the escort-type rehabilitation robot can guide reaching motion for the arm posture as well as the hand position and can also guide functional reach in collaboration with the user's trunk motion, proven by simulation and experiment.