Addressing data imbalance in Sim2Real: ImbalSim2Real scheme and its application in finger joint stiffness self-sensing for soft robot-assisted rehabilitation.

The simulation-to-reality (sim2real) problem is a common issue when deploying simulation-trained models to real-world scenarios, especially given the extremely high imbalance between simulation and real-world data (scarce real-world data). Although the cycle-consistent generative adversarial network (CycleGAN) has demonstrated promise in addressing some sim2real issues, it encounters limitations in situations of data imbalance due to the lower capacity of the discriminator and the indeterminacy of learned sim2real mapping. To overcome such problems, we proposed the imbalanced Sim2Real scheme (ImbalSim2Real). Differing from CycleGAN, the ImbalSim2Real scheme segments the dataset into paired and unpaired data for two-fold training. The unpaired data incorporated discriminator-enhanced samples to further squash the solution space of the discriminator, for enhancing the discriminator's ability. For paired data, a term targeted regression loss was integrated to ensure specific and quantitative mapping and further minimize the solution space of the generator. The ImbalSim2Real scheme was validated through numerical experiments, demonstrating its superiority over conventional sim2real methods. In addition, as an application of the proposed ImbalSim2Real scheme, we designed a finger joint stiffness self-sensing framework, where the validation loss for estimating real-world finger joint stiffness was reduced by roughly 41% compared to the supervised learning method that was trained with scarce real-world data and by 56% relative to the CycleGAN trained with the imbalanced dataset. Our proposed scheme and framework have potential applicability to bio-signal estimation when facing an imbalanced sim2real problem.

Modeling and mobile home monitoring of behavioral and psychological symptoms of dementia (BPSD).

With the increasing global aging population, dementia care has rapidly become a major social problem. Current diagnosis of Behavior and Psychological Symptoms of Dementia (BPSD) relies on clinical interviews, and behavioral rating scales based on a period of behavior observation, but these methods are not suitable for identification of occurrence of BPSD in the daily living, which is necessary for providing appropriate interventions for dementia, though, has been studied by few research groups in the literature. To address these issues, in this study developed a BPSD monitoring system consisting of a Psycho-Cognitive (PsyCo) BPSD model, a Behavior-Physio-Environment (BePhyEn) BPSD model, and an implementation platform. The PsyCo BPSD model provides BPSD assessment support to caregivers and care providers, while the BePhyEn BPSD model provides instantaneous alerts for BPSD enabled by a 24-hour home monitoring platform for early intervention, and thereby alleviation of burden to patients and caregivers. Data for acquiring the models were generated through extensive literature review and regularity determined. A mobile robot was utilized as the implementation platform for improving sensitivity of sensors for home monitoring, and elderly individual following algorithms were investigated. Experiments in a virtual home environment showed that, a virtual BPSD elderly individual can be followed safely by the robot, and BPSD occurrence could be identified accurately, demonstrating the possibility of modeling and identification of BPSD in home environment.

Corrigendum: Design of a SMA-based soft composite structure for wearable rehabilitation gloves.

[This corrects the article DOI: 10.3389/fnbot.2023.1047493.].

Design of a SMA-based soft composite structure for wearable rehabilitation gloves.

The combination of smart soft composite structure based shape memory alloy (SMA) and exoskeleton technology has the advantages of light weight, energy saving, and great human-exoskeleton interaction. However, there are no relevant studies on the application of SMA-based soft composite structure (SSCS) in hand exoskeletons. The main difficulty is that directional mechanical properties of SSCS need to comply with fingers movement, and SSCS can deliver enough output torque and displacement to the relevant joints. This paper aims to study the application of SSCS for wearable rehabilitation gloves and explore its bionic driving mechanism. This paper proposes a soft wearable glove (Glove-SSCS) for hand rehabilitation actuated by the SSCS, based on finger force analysis under different drive modes. The Glove-SSCS can support five-finger flexion and extension, weighs only 120 g, and adopts modular design. Each drive module adopts a soft composite structure. And the structure integrates actuation, sensing and execution, including an active layer (SMA spring), a passive layer (manganese steel sheet), a sensing layer (bending sensor) and connection layers. To obtain a high-performance SMA actuators, the performance of SMA materials was tested in terms of temperature and voltage, temperature at the shortest length, pre-tensile length and load. And the human-exoskeleton coupling model of Glove-SSCS is established and analyzed from force and motion. The results show that the Glove-SSCS can realize bidirectional movements of fingers flexion and extension, with ranges of motion are 90-110° and 30-40°, and their cycles are 13-19 s and 11-13 s. During the use of Glove-SSCS, the temperature of gloves is from 25 to 67°C, and the surface temperature of hands is from 32 to 36°C. The temperature of Glove-SSCS can be kept at the lowest temperature of SMA operation without much impact on the human body.

Design and kinematical performance analysis of the 7-DOF upper-limb exoskeleton toward improving human-robot interface in active and passive movement training.

BACKGROUND: Upper-limb rehabilitation robots have become an important piece of equipment in stroke rehabilitation. The design of exoskeleton mechanisms plays a key role to improve human-robot interface in the upper-limb movements under passive and active rehabilitation training. OBJECTIVE: This paper proposes a novel of the 7-DOF (RR-RR-PRR) under-actuated exoskeleton mechanism based on the characteristics of the upper-limb movements in both of active and passive training. This aim of the proposed work is to improve human-robot interface in rehabilitation training with robots. METHODS: Firstly, the characteristics of active and passive movement training are analyzed depending on the human upper-limb model. Then, a novel 7-DOF (RR-RR-PRR) exoskeleton mechanism is proposed based on the analyzed characteristics. After that, kinematical performances of the proposed exoskeleton are analyzed on the workspace, manipulability and manipulability ellipsoid by compared with the common exoskeleton configuration of the 7 DOFs (RRR-R-PRR) mechanism. In the end, the prototype is manufactured and tested by undergoing the experiments of single-joint passive movement training and multi-joint active movement training. The human-robot interface of the proposed exoskeleton is demonstrated by root mean square error, Pearson correlation coefficient, and the time-delay difference. RESULTS: The results of the kinematical performance show that the effective workspace and the flexibility of the exoskeleton with the proposed configuration are increased by 10.44% and 1.7%. In the single-joint passive movement training experiment, the root mean square errors are 6.986, 7.568, 5.846, and Pearson correlation coefficients are 0.989, 0.984, 0.988 at the shoulder joint and the elbow joint, respectively. The time-delay differences are not beyond 3.1%. In the multi-joint active movement training experiment, the root mean square errors are 9.312 and 7.677, and Pearson correlation coefficients are 0.906 and 0.968 at the shoulder joint and the elbow joint, respectively. The time-delay differences are not beyond 3.28%. CONCLUSIONS: The proposed 7 DOFs exoskeleton mechanism shows uniformity with that of the common exoskeleton on the same rehabilitation trajectory which is effective to improve human-robot interface under passive and active rehabilitation training.

Flexible lower limb exoskeleton systems: A review.

BACKGROUND: As an emerging exoskeleton robot technology, flexible lower limb exoskeleton (FLLE) integrates flexible drive and wearable mechanism, effectively solving many problems of traditional rigid lower limb exoskeleton (RLLE) such as higher quality, poorer compliance and relatively poor portability, and has become one of the important development directions in the field of active rehabilitation. OBJECTIVE: This review focused on the development and innovation process in the field of FLLE in the past decade. METHOD: Related literature published from 2010 to 2021 were searched in EI, IEEE Xplore, PubMed and Web of Science databases. Seventy target research articles were further screened and sorted through inclusion and exclusion criteria. RESULTS: FLLE is classified according to different driving modes, and the advantages and disadvantages of passive flexible lower limb exoskeletons and active flexible lower limb exoskeletons are comprehensively summarized. CONCLUSION: At present, FLLE's research is mainly based on cable drive, bionic pneumatic muscles followed and matured, and new exoskeleton designs based on smart material innovations also trend to diversify. In the future, the development direction of FLLE will be lightweight and drive compliance, and the multi-mode sensory feedback control theory, motion intention recognition theory and human-machine interaction theory will be combined to reduce the metabolic energy consumption of walking.

Human-robot coupling dynamic modeling and analysis for upper limb rehabilitation robots.

BACKGROUND: Upper limb rehabilitation robots have become an important piece of equipment in stroke rehabilitation. Human-robot coupling (HRC) dynamics play a key role in the control of rehabilitation robots to improve human-robot interaction. OBJECTIVE: This study aims to study the methods of modeling and analysis of HRC dynamics to realize more accurate dynamic control of upper limb rehabilitation robots. METHODS: By the analysis of force interaction between the human arm and the upper limb rehabilitation robot, the HRC torque is achieved by summing up the robot torque and the human arm torque. The HRC torque and robot torque of a 2-DOF upper limb rehabilitation robot (FLEXO-Arm) are solved by Lagrangian equation and step-by-step dynamic parameters identification method. RESULTS: The root mean square (RMS) is used to evaluate the accuracy of the HRC torque and the robot torque calculated by the parameter identification, and the error of both is about 10%. Moreover, the HRC torque and the robot torque are compared with the actual torque measured by torque sensors. The error of the robot torque is more than twice the HRC. Therefore, the HRC torque is more accurate than the actual torque. CONCLUSIONS: The proposed HRC dynamics effectively achieves more accurate dynamic control of upper limb rehabilitation robots.

Pilot Study of a Powered Exoskeleton for Upper Limb Rehabilitation Based on the Wheelchair.

To help hemiplegic patients with stroke to restore impaired or lost upper extremity functionalities efficiently, the design of upper limb rehabilitation robotics which can substitute human practice becomes more important. The aim of this work is to propose a powered exoskeleton for upper limb rehabilitation based on a wheelchair in order to increase the frequency of training and reduce the preparing time per training. This paper firstly analyzes the range of motion (ROM) of the flexion/extension, adduction/abduction, and internal/external of the shoulder joint, the flexion/extension of the elbow joint, the pronation/supination of the forearm, the flexion/extension and ulnar/radial of the wrist joint by measuring the normal people who are sitting on a wheelchair. Then, a six-degree-of-freedom exoskeleton based on a wheelchair is designed according to the defined range of motion. The kinematics model and workspace are analyzed to understand the position of the exoskeleton. In the end, the test of ROM of each joint has been done. The maximum error of measured and desired shoulder flexion and extension joint angle is 14.98%. The maximum error of measured and desired elbow flexion and extension joint angle is 14.56%. It is acceptable for rehabilitation training. Meanwhile, the movement of drinking water can be realized in accordance with the range of motion. It demonstrates that the proposed upper limb exoskeleton can also assist people with upper limb disorder to deal with activities of daily living. The feasibility of the proposed powered exoskeleton for upper limb rehabilitation training and function compensating based on a wheelchair is proved.