The pattern recognition (PR) based on surface electromyography (sEMG) could improve the quality of daily life of amputees. However, the lack of robustness and adaptability hinders its practical application. To realize the long-term reliability and user adaptability simultaneously, a novel multi-task dual-stream supervised domain adaptation (MDSDA) network based on convolutional neural network (CNN) was proposed. A long-term multi-subject sEMG signal acquisition was conducted to validate the performance of MDSDA, recruiting 12 able-bodied subjects. A total of thirty gestures were used for the acquisition, including one set of static gestures and two sets of dynamic gestures. The longterm multi-subject sEMG dataset is publicly available at the website (https://github.com/tinker1017/LONG-TERM-ANDMULTI-SUBJECT-SEMG-DATASET). Four train-test estimations were designed to evaluate the robustness and adaptability of MDSDA. The results showed that MDSDA outperformed CNN and fune-tuning. Furthermore, we studied the divisibility between static and dynamic gestures that performed similar actions. The outcomes demonstrated that there existed high separability between them. This may be helpful to reduce the signal collection burden. Experimental results proved MDSDA has the potential to provide a robust and generalized PR system for the clinic applications.

Rehabilitation robots with intent recognition are helping people with dysfunction to enjoy better lives. Many rehabilitation robots with intent recognition have been developed by academic instituti... As a technology, intent recognition can become a part of rehabilitation, assist patients to complete daily life activities, and improve their quality of life.Rehabilitation training equipment for t...

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.

The capability to generate electricity from human motion can reduce the battery requirements for wearable devices. The key challenge faced by wearable energy harvesters is the generation of sufficient power without interfering with the wearer. To our knowledge, currently available joint-motion energy harvesters are all for level walking (LW). In this study, we developed an energy-harvesting exoskeleton that can generate electricity during LW, downhill walking (DW), and stair descent (SD) without sensors. Bi-directional knee motion is transformed to the unidirectional rotation of the generator by one-way bearing and appropriate gear train. The average electricity generated by each leg for the three modes of walking is $5.4 \pm 0.8{\rm{\ W\ }}({\text{LW},\ 5{\rm{\ km}}/\mathrm{h}})$ , $6.5 \pm 0.6{\rm{W\ }}({\text{DW},\ 5{\rm{\ km}}/\mathrm{h}})$ , and $8.2 \pm 0.4{\rm{W\ }}({\text{SD},4{\rm{\ stairs}}/\mathrm{s}})$ , respectively. The Pearson coefficients of the knee angle under exo-on and no-exo conditions are 0.995 (LW), 0.996 (DW), and 0.999 (SD), respectively. The cost of harvesting is −0.006 (LW) and −0.01 (DW). The increase in the metabolic rate in energy harvesting on condition is merely 2.3% (LW) and 1.6% (DW) compared to the no-exo condition. There is no significant increase in the metabolic cost while generating electricity $(P\ = \ 0.363,$ LW; $P = 0.662$ , DW). In summary, the knee energy-harvesting exoskeleton developed in this study can generate substantial and sustainable electricity with little extra effort, which indicates its potential as a means to charge-powered portable exoskeletons and prosthetics.

Model-based impedance learning control can provide variable impedance regulation for robots through online impedance learning without interaction force sensing. However, the existing related results only guarantee the closed-loop control systems to be uniformly ultimately bounded (UUB) and require the human impedance profiles being periodic, iteration-dependent, or slowly varying. In this article, a repetitive impedance learning control approach is proposed for physical human–robot interaction (PHRI) in repetitive tasks. The proposed control is composed of a proportional-differential (PD) control term, an adaptive control term, and a repetitive impedance learning term. Differential adaptation with projection modification is designed for estimating robotic parameters uncertainties in the time domain, while fully saturated repetitive learning is proposed for estimating time-varying human impedance uncertainties in the iterative domain. Uniform convergence of tracking errors is guaranteed by the PD control and the use of projection and full saturation in the uncertainties estimation and is theoretically proved based on a Lyapunov-like analysis. In impedance profiles, the stiffness and damping are composed of an iteration-independent term and an iteration-dependent disturbance, which are estimated by repetitive learning and compressed by the PD control, respectively. Therefore, the developed approach can be applied to the PHRI where iteration-dependent disturbances exist in the stiffness and damping. The control effectiveness and advantages are validated by simulations on a parallel robot in a repetitive following task.

Isokinetic muscle strength training is a type of rehabilitation therapy that can effectively improve the motor dysfunction of limbs and enhance joint muscle strength. In active isokinetic training, how to plan the speed reasonably and improve the performance of speed control to fit individualized needs is the core problem of isokinetic muscle strength training equipment control. In this study, a control strategy for isokinetic muscle strength training equipment is proposed, which identifies the user’s motion intention and performs speed planning according to the user’s motion participation so that the equipment produces a dynamic damping effect. This ensures active interaction between the user and the equipment and meets the force-speed requirements of elbow joint muscle training. To handle the uncertain disturbance in the training process, a linear extended state observer is used to estimate the disturbance term, and a tracking differentiator and nonlinear error feedback control are used to improve the speed tracking performance. The experimental results show that the proposed control scheme realizes effective speed control in isokinetic muscle strength training. Note to Practitioners—The bottleneck of isokinetic muscle strength training equipment control system lies in how to plan the speed according to the patient’s movement intention reasonably, and how to improve the performance of speed control in the process of rehabilitation training. This paper designs a speed auto disturbance rejection control strategy for isokinetic muscle strength training system with external interference and model uncertainty in practical application. The experimental results show that the performance of the controller is better than the traditional PI controller, and it can realize isokinetic muscle training. This study provides a method for practitioners interested in developing control systems for isokinetic rehabilitation training equipment.