As an important movement of the daily living activities, sit-to-stand (STS) movement is usually a difficult task facing elderly and dependent people. In this article, a novel impedance modulation strategy of a lower-limb exoskeleton is proposed to provide appropriate power and balance assistance during STS movements while preserving the wearer’s control priority. The impedance modulation control strategy ensures adaptation of the mechanical impedance of the human–exoskeleton system toward a desired one requiring less wearer’s effect while reinforcing the wearer’s balance control ability during STS movements. A human joint torque observer is designed to estimate the joint torques developed by the wearer using joint position kinematics instead of electromyography or force sensors; a time-varying desired impedance model is proposed according to the wearer’s lower-limb motion ability. A virtual environmental force is designed for balance reinforcement control. Stability and robustness of the proposed method are theoretically analyzed. Simulations are implemented to illustrate the characteristics and performance of the proposed approach. Experiments with four healthy subjects are carried out to evaluate the effectiveness of the proposed method and show satisfactory results in terms of appropriate power assist and balance reinforcement.

Parkinson’s disease (PD) is the second most common neurodegenerative disease affecting millions worldwide. Bespoke subject-specific treatment (medication or deep brain stimulation (DBS)) is critical for management, yet depends on precise assessment cardinal PD symptoms -bradykinesia, rigidity and tremor. Clinician diagnosis is the basis of treatment, yet it allows only a cross-sectional assessment of symptoms which can vary on an hourly basis and is liable to inter-and intra-rater subjectivity across human examiners. Automated symptomatic assessment has attracted significant interest to optimise treatment regimens between clinician visits, however, no wearable has the capacity to simultaneously assess all three cardinal symptoms. Challenges in the measurement of rigidity, mapping muscle activity out-of-clinic and sensor fusion have inhibited translation. In this study, we address all through a novel wearable sensor system and learning algorithms. The sensor system is composed of a force-sensor, three inertial measurement units (IMUs) and four custom mechanomyography (MMG) sensors. The system was tested in its capacity to predict Unified Parkinson’s Disease Rating Scale (UPDRS) scores based on quantitative assessment of bradykinesia, rigidity and tremor in PD patients. 23 PD patients were tested with the sensor system in parallel with exams conducted by treating clinicians and 10 healthy subjects were recruited as a comparison control group. Results prove the system accurately predicts UPDRS scores for all symptoms (85.4% match on average with physician assessment) and discriminates between healthy subjects and PD patients (96.6% on average). MMG features can also be used for remote monitoring of severity and fluctuations in PD symptoms out-of-clinic. This closed-loop feedback system enables individually tailored and regularly updated treatment, facilitating better outcomes for a very large patient population.

Reducing human lower limb effort during walking is still a challenge for wearable exoskeletons. From the wearer's viewpoint, joint impedance increases while wearing the exoskeleton due to the fact that the exoskeleton's mass, friction, and gravity are added to those of the human lower limbs. Active impedance is required from the exoskeleton to compensate not only the exoskeleton's mechanical impedance but also that of the human limb. This paper presents a nonlinear impedance reduction control (IRC) approach applied to a knee joint exoskeleton during flexion/extension movements. The IRC approach ensures impedance adaptation of the human-exoskeleton system toward a desired level. The proposed method is an alternative to sensor-based impedance controllers where the human joint torque is estimated using electromyography (EMG) sensors, force/torque sensors, and so on. The performance of the proposed approach and its robustness with respect to modeling uncertainties are theoretically analyzed and discussed. Experiments were conducted with four healthy subjects to verify the effectiveness of human effort reduction. The root mean square of the EMG signals of the muscles involved in the flexion/extension was used as a metric. The experimental results show that EMG signal levels can be efficiently reduced by inertia, damping, and gravity compensation compared with those obtained without wearing the exoskeleton.

Human-robot cooperation is vital for optimising powered assist of lower limb exoskeletons (LLEs). Robotic capacity to intelligently adapt to human force, however, demands a fusion of data from exoskeleton and user state for smooth human-robot synergy. Muscle activity, mapped through electromyography (EMG) or mechanomyography (MMG) is widely acknowledged as usable sensor input that precedes the onset of human joint torque. However, competing and complementary information between such physiological feedback is yet to be exploited, or even assessed, for predictive LLE control. We investigate complementary and competing benefits of EMG and MMG sensing modalities as a means of calculating human torque input for assist-as-needed (AAN) LLE control. Three biomechanically agnostic machine learning approaches, linear regression, polynomial regression, and neural networks, are implemented for joint torque prediction during human-exoskeleton interaction experiments. Results demonstrate MMG predicts human joint torque with slightly lower accuracy than EMG for isometric human-exoskeleton interaction. Performance is comparable for dynamic exercise. Neural network models achieve the best performance for both MMG and EMG (94.8±0.7% with MMG and 97.6±0.8% with EMG (Mean ± SD)) at the expense of training time and implementation complexity. This investigation represents the first MMG human joint torque models for LLEs and their first comparison with EMG. We provide our implementations for future investigations (https://github.com/cic12/ieee_appx).

In this article, a proxy-based force control method is proposed for three important human–robot interaction modes: zero-impedance mode, force assistive mode, and large force mode. A two-mass dynamic-model-based nonlinear disturbance observer is used to meet the zero-impedance output and accurate force tracking requirements with respect to disturbances from the wearer and environment. Additionally, significant force compliance can be achieved to guarantee the wearer’s safety when the interaction torque is large. The proposed method is evaluated via experiments by comparison to the conventional proportional–integral–derivative and proxy-based sliding mode control methods. The results indicate that the proposed approach achieves better force tracking accuracy, robustness, and force compliance in three-mode human–robot interactions.

Loss of mobility and/or balance resulting from neural trauma is a critical public health issue. Robotic exoskeletons hold great potential for rehabilitation and assisted movement. However, the synergy of robot operation with human effort remains a problem. In particular, optimal assist-as-needed (AAN) control remains unresolved given pathological variance among patients. We introduce a model predictive control (MPC) architecture for lower limb exoskeletons that achieves on-the-fly transitions between modes of assistance. The architecture implements a fuzzy logic algorithm (FLA) to map key modes of assistance based on human involvement. Three modes are utilised: passive, for human relaxed and robot dominant; active-assist, for human cooperation with the task; and safety, in the case of human resistance to the robot. Electromyography (EMG) signals are further employed to predict the human torque. EMG output is used by the MPC for trajectory following and by the FLA for decision making. Experimental validation using a 1-DOF knee exoskeleton demonstrates the controller tracking a sinusoidal trajectory with relaxed, assistive, and resistive operational modes. Results demonstrate rapid and appropriate transfers among the assistance modes, and satisfactory AAN performance in each case, offering a new level of human-robot synergy for mobility assist and rehabilitation.