Adaptive interaction torque-based AAN control for lower limb rehabilitation exoskeleton

Finite-time compensation control with dead-zone estimation for a rehabilitative walker considering internal disturbance forces

This study discusses a finite-time compensation tracking control method for a rehabilitative training walker. The dynamic model with input dead zone was constructed to describe the walker, and a finite-time disturbance forces observation method was proposed based on the impact mechanism on tracking performance. This approach is novel in that the disturbance forces were observed in reverse through their effects on tracking performance, thus successfully obtaining the disturbance forces of the walker. To ensure the practical finite-time stability of the system, the nonlinear finite-time compensation tracking controller with stochastic configuration networks (SCN) dead-zone estimation was built for the rehabilitative walker. Simulation results and comparative analyses confirmed that the proposed compensation control method effectively restrains dead zone and internal disturbance forces.

Backstepping control based on adaptive neural network and disturbance observer for reconfigurable variable stiffness actuator

Reconfigurable variable stiffness actuator (RVSA) has attracted increasing attention in robotics due to its safety, compliance, and robustness. However, the control of the RVSA is challenging due to nonlinear factors such as high-order nonlinear dynamic, model uncertainties, time-varying model parameters, and disturbances. In this paper, firstly, a lightweight RVSA structure with both passive and active nonlinear variable stiffness characteristic is developed. Secondly, a dynamic surface backstepping control method based on a radial basis neural network and disturbance observer (DSBC-RBFNN-DOB) is proposed to achieve position control of the lightweight RVSA with matched and unmatched uncertainties. To address solve the "complexity explosion" and noise problems in traditional backstepping control, the dynamic surface backstepping control (DSBC) method is used to design the controller. Then, a method based on radial basis neural network (RBFNN) and disturbance observer (DOB) are used to compensate for the matched and unmatched uncertainties in the link and motor. In this method, the matched uncertainties are compensated using RBFNN, and the DOB is integrated to compensate RBFNN approximation errors and unmatched uncertainties. Through Lyapunov stability analysis, the semi-global boundedness of the controller is proven. Finally, the proposed method is simulated and actually implemented, verifying the effectiveness of the method. Simulation and experimental results show that the root mean square error (RMSE) of the proposed method is only 0.97277° and 0.6418°, respectively. Compared with PID, DSBC, and DSBC-RBFNN, the error reduction percentages in simulation (experiment) are 85.6 % (88.9 %), 49.4 % (88.4 %) and 36.1 % (80.0 %) respectively.

A Self-Coordinating Controller with Balance-Guiding Ability for Lower-Limb Rehabilitation Exoskeleton Robot

The restricted posture and unrestricted compliance brought by the controller during human-exoskeleton interaction (HEI) can cause patients to lose balance or even fall. In this article, a self-coordinated velocity vector (SCVV) double-layer controller with balance-guiding ability was developed for a lower-limb rehabilitation exoskeleton robot (LLRER). In the outer loop, an adaptive trajectory generator that follows the gait cycle was devised to generate a harmonious hip-knee reference trajectory on the non-time-varying (NTV) phase space. In the inner loop, velocity control was adopted. By searching the minimum L2 norm between the reference phase trajectory and the current configuration, the desired velocity vectors in which encouraged and corrected effects can be self-coordinated according to the L2 norm were obtained. In addition, the controller was simulated using an electromechanical coupling model, and relevant experiments were carried out with a self-developed exoskeleton device. Both simulations and experiments validated the effectiveness of the controller.

Evaluation of a machine-learning-driven active-passive upper-limb exoskeleton robot: Experimental human-in-the-loop study

Evaluating exoskeleton actuation methods and designing an effective controller for these exoskeletons are both challenging and time-consuming tasks. This is largely due to the complicated human-robot interactions, the selection of sensors and actuators, electrical/command connection issues, and communication delays. In this research, a test framework for evaluating a new active-passive shoulder exoskeleton was developed, and a surface electromyography (sEMG)-based human-robot cooperative control method was created to execute the wearer's movement intentions. The hierarchical control used sEMG-based intention estimation, mid-level strength regulation, and low-level actuator control. It was then applied to shoulder joint elevation experiments to verify the exoskeleton controller's effectiveness. The active-passive assistance was compared with fully passive and fully active exoskeleton control using the following criteria: (1) post-test survey, (2) load tolerance duration, and (3) computed human torque, power, and metabolic energy expenditure using sEMG signals and inverse dynamic simulation. The experimental outcomes showed that active-passive exoskeletons required less muscular activation torque (50%) from the user and reduced fatigue duration indicators by a factor of 3, compared to fully passive ones.

Fuzzy Adaptive Passive Control Strategy Design for Upper-Limb End-Effector Rehabilitation Robot

Robot-assisted rehabilitation therapy has been proven to effectively improve upper-limb motor function in stroke patients. However, most current rehabilitation robotic controllers will provide too much assistance force and focus only on the patient's position tracking performance while ignoring the patient's interactive force situation, resulting in the inability to accurately assess the patient's true motor intention and difficulty stimulating the patient's initiative, thus negatively affecting the patient's rehabilitation outcome. Therefore, this paper proposes a fuzzy adaptive passive (FAP) control strategy based on subjects' task performance and impulse. To ensure the safety of subjects, a passive controller based on the potential field is designed to guide and assist patients in their movements, and the stability of the controller is demonstrated in a passive formalism. Then, using the subject's task performance and impulse as evaluation indicators, fuzzy logic rules were designed and used as an evaluation algorithm to quantitively assess the subject's motor ability and to adaptively modify the stiffness coefficient of the potential field and thus change the magnitude of the assistance force to stimulate the subject's initiative. Through experiments, this control strategy has been shown to not only improve the subject's initiative during the training process and ensure their safety during training but also enhance the subject's motor learning ability.

Model-Based Mid-Level Regulation for Assist-As-Needed Hierarchical Control of Wearable Robots: A Computational Study of Human-Robot Adaptation

The closed-loop human–robot system requires developing an effective robotic controller that considers models of both the human and the robot, as well as human adaptation to the robot. This paper develops a mid-level controller providing assist-as-needed (AAN) policies in a hierarchical control setting using two novel methods: model-based and fuzzy logic rule. The goal of AAN is to provide the required extra torque because of the robot’s dynamics and external load compared to the human limb free movement. The human–robot adaptation is simulated using a nonlinear model predictive controller (NMPC) as the human central nervous system (CNS) for three conditions of initial (the initial session of wearing the robot, without any previous experience), short-term (the entire first session, e.g., 45 min), and long-term experiences. The results showed that the two methods (model-based and fuzzy logic) outperform the traditional proportional method in providing AAN by considering distinctive human and robot models. Additionally, the CNS actuator model has difficulty in the initial experience and activates both antagonist and agonist muscles to reduce movement oscillations. In the long-term experience, the simulation shows no oscillation when the CNS NMPC learns the robot model and modifies its weights to simulate realistic human behavior. We found that the desired strength of the robot should be increased gradually to ignore unexpected human–robot interactions (e.g., robot vibration, human spasticity). The proposed mid-level controllers can be used for wearable assistive devices, exoskeletons, and rehabilitation robots.