Reduced Adaptive Fuzzy Decoupling Control for Lower Limb Exoskeleton

Adaptive neural fault-tolerant prescribed performance control of a rehabilitation exoskeleton for lower limb passive training

This article studies the passive tracking problem of a wearable exoskeleton for lower limb rehabilitation therapy in the face of unmodeled dynamics, interactive friction, disturbance, prescribed performance constraints, and actuator faults. Adaptive neural networks and a smooth performance function are incorporated to establish a novel fault-tolerant tracking scheme, which can not only compensate for the nonlinear uncertainties and disturbance, but also handle the actuator fault with guaranteed tracking performance. A state feedback controller is presented by using the full state information and an output feedback controller is developed when the angular velocity is unavailable. The differential explosion issue of the backstepping technique is resolved by constructing a first-order filter and the unmeasurable velocity is estimated by a nonlinear observer. Semiglobal uniform boundedness stabilities of the exoskeleton system are proved via the Lyapunov direct method. The tracking performances of the designed control approaches are tested by comparative simulations.

A multimodal framework based on deep belief network for human locomotion intent prediction

Accurate prediction of human locomotion intent benefits the seamless switching of lower limb exoskeleton controllers in different terrains to assist humans in walking safely. In this paper, a deep belief network (DBN) was developed to construct a multimodal framework for recognizing various locomotion modes and predicting transition tasks. Three fusion strategies (data level, feature level, and decision level) were explored, and optimal network performance was obtained. This method could be tested on public datasets. For the continuous performance of steady state, the best prediction accuracy achieved was 97.64% in user-dependent testing and 96.80% in user-independent testing. During the transition state, the system accurately predicted all transitions (user-dependent: 96.37%, user-independent: 95.01%). The multimodal framework based on DBN can accurately predict the human locomotion intent. The experimental results demonstrate the potential of the proposed model in the volition control of the lower limb exoskeleton.

Active Human-Following Control of an Exoskeleton Robot With Body Weight Support

This article presents an active human-following control of the lower limb exoskeleton for gait training. First, to improve safety, considering the human balance, the OpenPose-based visual feedback is used to estimate the individual's pose, then, the active human-following algorithm is proposed for the exoskeleton robot to achieve the body weight support and active human-following. Second, taking the human's intention and voluntary efforts into account, we develop a long short-term memory (LSTM) network to extract surface electromyography (sEMG) to build the estimation model of joints' angles, that is, the multichannel sEMG signals can be correlated with flexion/extension (FE) joints' angles of the human lower limb. Finally, to make the robot motion adapt to the locomotion of subjects under uncertain nonlinear dynamics, an adaptive control strategy is designed to drive the exoskeleton robot to track the desired locomotion trajectories stably. To verify the effectiveness of the proposed control framework, several recruited subjects participated in the experiments. Experimental results show that the proposed joints' angles estimation model based on the LSTM network has a higher estimation accuracy and predicted performance compared with the existing deep neural network, and good simultaneous locomotion tracking performance is achieved by the designed control strategy, which indicates that the proposed control can assist subjects to perform gait training effectively.

Computed torque control and force analysis for mechanical leg with variable rotation axis powered by servo pneumatic muscle

It is difficult for a humanoid leg driven by two groups of antagonistic pneumatic muscles (PMs) to achieve a flexible humanoid gait, and its inherent strong coupling nonlinear characteristics make it hard to achieve good tracking performance in a large range of motion. Therefore, a four-bar linkage bionic knee joint structure with a variable axis and a double closed-loop servo position control strategy based on computed torque control are designed to improve anthropomorphic characteristics and the dynamic performance of the bionic mechanical leg powered by servo pneumatic muscle (SPM). Firstly, the relationship between the joint torque, the initial jump angle and the bounce height of the mechanical leg is established, and then we design a double-joint PM bionic mechanical leg containing a four-bar linkage mechanism of the knee joint. Secondly, a cascade position control strategy is developed, which consists of the outer position loop and the inner contraction force loop, and the mapping relationship is designed between joint torque and antagonistic PM contraction force. Finally, we further project bounce action timing of mechanical leg to realize the periodic jumping movement of the mechanical leg, and simulation and physical experiments of the real-style machine platform have been provided to demonstrate the effectiveness of the designed SPM controller.

Review of adaptive control for stroke lower limb exoskeleton rehabilitation robot based on motion intention recognition

Stroke is a significant cause of disability worldwide, and stroke survivors often experience severe motor impairments. Lower limb rehabilitation exoskeleton robots provide support and balance for stroke survivors and assist them in performing rehabilitation training tasks, which can effectively improve their quality of life during the later stages of stroke recovery. Lower limb rehabilitation exoskeleton robots have become a hot topic in rehabilitation therapy research. This review introduces traditional rehabilitation assessment methods, explores the possibility of lower limb exoskeleton robots combining sensors and electrophysiological signals to assess stroke survivors' rehabilitation objectively, summarizes standard human-robot coupling models of lower limb rehabilitation exoskeleton robots in recent years, and critically introduces adaptive control models based on motion intent recognition for lower limb exoskeleton robots. This provides new design ideas for the future combination of lower limb rehabilitation exoskeleton robots with rehabilitation assessment, motion assistance, rehabilitation treatment, and adaptive control, making the rehabilitation assessment process more objective and addressing the shortage of rehabilitation therapists to some extent. Finally, the article discusses the current limitations of adaptive control of lower limb rehabilitation exoskeleton robots for stroke survivors and proposes new research directions.

Effects of Mobile Phone Use on Gait and Balance Control in Young Adults: A Hip-Ankle Strategy

BACKGROUND

This study aimed to derive the effects of walking while using a mobile phone on balance perturbation and joint movement among young adults.

METHODS

Sixteen healthy college students with no history of brain injury were tested. The participants were asked to walk under four different conditions: (1) walking, (2) browsing, (3) dialing, and (4) texting. Indicators related to balance control and lower limb kinematic/kinetic parameters were analyzed using the continuous relative phase and statistical nonparametric mapping methods.

RESULTS

Walking while using a mobile phone slowed participants' gait speed and reduced the cadence, stride length, and step length. The posterior tilt angle (0-14%, 57-99%), torque of the hip flexion (0-15%, 30-35%, 75-100%), and angle of the hip flexion (0-28%, 44-100%) decreased significantly. The activation of biceps femoris and gastrocnemius, hip stiffness, and ankle stiffness increased significantly. This impact on gait significantly differed among three dual tasks: texting > browsing > dialing.

CONCLUSION

Che overlap of walking and mobile phone use affects the gait significantly. The "hip-ankle strategy" may result in a "smooth" but slower gait, while this strategy was deliberate and tense. In addition, this adjustment also increases the stiffness of the hip and ankle, increasing the risk of fatigue. Findings regarding this effect may prove that even for young healthy adults, walking with mobile phone use induces measurable adjustment of the motor pattern. These results suggest the importance of simplifying the control of the movement.

Tracking approach of double pendulum cranes with variable rope lengths using sliding mode technique

Cranes are widely employed for transportation in many industrial fields. However, strong couplings and high nonlinearities increase difficulty of designing effective control methods for crane systems. Additionally, complex environments may bring uncertainties and unfavorable factors for the control problem. In this paper, we design an effective tracking approach for double pendulum cranes considering variable rope lengths problem by using sliding mode technique, to further improve the robustness and achieve the control objectives of accurate tracking, while effective suppression for double pendulum swing is also ensured. We design a proper sliding mode surface containing all state variables, which effectively suppresses swing angles of the payload and the hook. Then, the time delay estimator technique is used to estimate parameter uncertainty-related terms and the estimation errors are effectively tackled by using sliding mode technique. The effectiveness is rigorously proved by Lyapunov stability theory. At last, comprehensive simulations are implemented to show the proper performance.

Fixed-Time Adaptive Neural Network Control for Nonlinear Systems With Input Saturation

This study concentrates on the tracking control problem for nonlinear systems subject to actuator saturation. To improve the performance of the controller, we propose a fixed-time tracking control scheme, in which the upper bound of the convergence time is independent of the initial conditions. In the control scheme, first, a smooth nonlinear function is employed to approximate the saturation function so that the controller can be designed under the framework of backstepping. Then, the effect of input saturation is compensated by introducing an auxiliary system. Furthermore, a fixed-time adaptive neural network control method is given with the help of fixed-time control theory, in which the dynamic order of controllers is reduced to a certain extent since there is only one updating law in the entire control design. Through rigorous theoretical analysis, it is concluded that the proposed control scheme can guarantee that: 1) the output tracking error can converge to a small neighborhood near the origin in a fixed time and 2) all signals in the closed-loop system are bounded. Finally, a numerical example and a practical example based on the single-link manipulator are provided to verify the effectiveness of the proposed method.

Robust walking control of a lower limb rehabilitation exoskeleton coupled with a musculoskeletal model via deep reinforcement learning

BACKGROUND

Few studies have systematically investigated robust controllers for lower limb rehabilitation exoskeletons (LLREs) that can safely and effectively assist users with a variety of neuromuscular disorders to walk with full autonomy. One of the key challenges for developing such a robust controller is to handle different degrees of uncertain human-exoskeleton interaction forces from the patients. Consequently, conventional walking controllers either are patient-condition specific or involve tuning of many control parameters, which could behave unreliably and even fail to maintain balance.

METHODS

We present a novel, deep neural network, reinforcement learning-based robust controller for a LLRE based on a decoupled offline human-exoskeleton simulation training with three independent networks, which aims to provide reliable walking assistance against various and uncertain human-exoskeleton interaction forces. The exoskeleton controller is driven by a neural network control policy that acts on a stream of the LLRE's proprioceptive signals, including joint kinematic states, and subsequently predicts real-time position control targets for the actuated joints. To handle uncertain human interaction forces, the control policy is trained intentionally with an integrated human musculoskeletal model and realistic human-exoskeleton interaction forces. Two other neural networks are connected with the control policy network to predict the interaction forces and muscle coordination. To further increase the robustness of the control policy to different human conditions, we employ domain randomization during training that includes not only randomization of exoskeleton dynamics properties but, more importantly, randomization of human muscle strength to simulate the variability of the patient's disability. Through this decoupled deep reinforcement learning framework, the trained controller of LLREs is able to provide reliable walking assistance to patients with different degrees of neuromuscular disorders without any control parameter tuning.

RESULTS AND CONCLUSION

A universal, RL-based walking controller is trained and virtually tested on a LLRE system to verify its effectiveness and robustness in assisting users with different disabilities such as passive muscles (quadriplegic), muscle weakness, or hemiplegic conditions without any control parameter tuning. Analysis of the RMSE for joint tracking, CoP-based stability, and gait symmetry shows the effectiveness of the controller. An ablation study also demonstrates the strong robustness of the control policy under large exoskeleton dynamic property ranges and various human-exoskeleton interaction forces. The decoupled network structure allows us to isolate the LLRE control policy network for testing and sim-to-real transfer since it uses only proprioception information of the LLRE (joint sensory state) as the input. Furthermore, the controller is shown to be able to handle different patient conditions without the need for patient-specific control parameter tuning.

A Stackelberg Game-Theoretic Exploration Rendering Robustness and Optimality for Performance Improvement of Fuzzy Mechanical Systems

We consider mechanical systems with uncertainty. The uncertainty may be time varying. The bound of the uncertainty is described by its fuzzy characteristics. To design a feasible control, we start with a robust phase, which renders a control scheme that guarantees the system performance regardless of the actual value of the uncertainty. This robust phase is then followed by an optimal phase. There are design parameters in the control, which can be fine-tuned. We proposed multiple performance objectives. The goal of the choice of the control design parameters is to minimize the performance objectives. However, since these objectives are nonconciliating (meaning one's minimum is not the other one's minimum), we invoke the Stackelberg strategy for the optimal parameters. The game strategy mimics two players: one is the leader and one is the follower. Through the interplay between the two players, we show how to select the design parameters. The design procedure in both robust and optimal phases is demonstrated by a coupled inverted pendulum system.

Adaptive Swarm Fuzzy Logic Controller of Multi-Joint Lower Limb Assistive Robot

The idea of developing a multi-joint rehabilitation robot is to satisfy the demands for recovery of lower limb functionality in hemiplegic impairments and assist the physiotherapists with their therapy plans. This work aims at to implement the Lyapunov Adaptive and Swarm-Fuzzy Logic Control (LASFC) strategy of 4-degree of freedom (4-DoF) Lower Limb Assistive Robot (LLAR) application, in which the control law is an integration of swarm-fuzzy logic control (SFLC) and Lyapunov adaptive control (LAC) with particle swarm optimization (PSO). The controller is established based on the sliding filtered steady-state error for SFLC. Its parameters are tuned by using PSO for the mathematical model of LLAR. The fuzzy defuzzification membership is set based on the tuned parameters for the real-time control system. LAC strategy is determined using stability analysis of the system to choose the controller’s parameters by observation of the system’s output and reference. The control law implemented in LLAR is the integration of SFLC and LAC to adjust the input voltage of joints. The parameters tuned by PSO are compared with the genetic algorithm (GA) statistically. In addition, the real-time trajectory tracking of the proposed controller for each joint is compared with LAC and SFLC separately. The experiment revealed that the LASFC has superior performance to the other two methods in trajectory tracking. For example, the average error for left hip by LASFC is 53.57% and 68% lower than SFLC and LAC, respectively. By the statistical analysis, it can be ascertained that the LASFC strategy performed efficiently for real-time control of the joint trajectory tracking.

Robust LQR-Based Neural-Fuzzy Tracking Control for a Lower Limb Exoskeleton System with Parametric Uncertainties and External Disturbances

The design of an accurate control scheme for a lower limb exoskeleton system has few challenges due to the uncertain dynamics and the unintended subject's reflexes during gait rehabilitation. In this work, a robust linear quadratic regulator- (LQR-) based neural-fuzzy (NF) control scheme is proposed to address the effect of payload uncertainties and external disturbances during passive-assist gait training. Initially, the Euler-Lagrange principle-based nonlinear dynamic relations are established for the coupled system. The input-output feedback linearization approach is used to transform the nonlinear relations into a linearized state-space form. The architecture of the adaptive neuro-fuzzy inference system (ANFIS) and used membership function are briefly explained. While varying mass parameters up to 20%, three robust neural-fuzzy datasets are formulated offline with the joint error vector and LQR control input. Thereafter, to deal with external interferences, an error dynamics with a disturbance estimator is presented using an online adaptation of the firing strength matrix. The Lyapunov theory is carried out to ensure the asymptotic stability of the coupled human-exoskeleton system in view of the proposed controller. The gait tracking results for the proposed control scheme (RLQR-NF) are presented and compared with the exponential reaching law-based sliding mode (ERL-SM) controller. Furthermore, to investigate the robustness of the proposed control over LQR control, a comparative performance analysis is presented for two cases of parametric uncertainties and external disturbances. The first case considers the 20% raise in mass values with a trigonometric form of disturbances, and the second case includes the effect of the 30% increment in mass values with a random form of disturbances. The simulation runs have shown the promising gait tracking aspects of the designed controller for passive-assist gait training.

A New Switching Adaptive Fuzzy Controller with an Application to Vibration Control of a Vehicle Seat Suspension Subjected to Disturbances

This paper proposes a new switching adaptive fuzzy controller and applies it to vibration control of a vehicle seat suspension equipped with a semi-active magnetorheological (MR) damper. The proposed control system consists of three functioned filters: (1) Filter 1: a model of interval type 2 fuzzy to compensate disturbances; (2) Filter 2: a ‘switching term’ to evaluate the magnitude of disturbance; and (3) Filter 3: a group of adaptation laws to enhance the robustness of control input. These filters play a role of powerful shields to improve control performance and guarantee the stability of the applied system subjected to external disturbances. After embedding a PID (proportional-integral-derivative) model into Riccati-like equation, main control parameters are updated based on the adaptation laws. The proposed controller is then synthesized in two different cases: high disturbance and small disturbance. For the high disturbance, a special type of sliding surface function, which relates to an exponential function and its t-norm, is used to increase the energy of control system. For the small disturbance, the energy from the modified t-norm of the sliding surface is neglected to reduce the energy consumption with maintaining the desired performance. To demonstrate the effectiveness of the proposed controller, a vehicle seat suspension installed with controllable MR damper is adopted to reflect the robustness against external disturbances corresponding to road excitations. It is validated from computer simulation that the proposed controller can provide better vibration control performance than other existing robust controllers showing excellent control stability with well-reduced displacement and velocity at the position of the seat.

Lower Limb Kinematics Trajectory Prediction Using Long Short-Term Memory Neural Networks

This study determined whether the kinematics of lower limb trajectories during walking could be extrapolated using long short-term memory (LSTM) neural networks. It was hypothesised that LSTM auto encoders could reliably forecast multiple time-step trajectories of the lower limb kinematics, specifically linear acceleration (LA) and angular velocity (AV). Using 3D motion capture, lower limb position-time coordinates were sampled (100 Hz) from six male participants (age 22 ± 2 years, height 1.77 ± 0.02 m, body mass 82 ± 4 kg) who walked for 10 min at 5 km/h on a 0% gradient motor-driven treadmill. These data were fed into an LSTM model with a sliding window of four kinematic variables with 25 samples or time steps: LA and AV for thigh and shank. The LSTM was tested to forecast five samples (i.e., time steps) of the four kinematic input variables. To attain generalisation, the model was trained on a dataset of 2,665 strides from five participants and evaluated on a test set of 1 stride from a sixth participant. The LSTM model learned the lower limb kinematic trajectories using the training samples and tested for generalisation across participants. The forecasting horizon suggested higher model reliability in predicting earlier future trajectories. The mean absolute error (MAE) was evaluated on each variable across the single tested stride, and for the five-sample forecast, it obtained 0.047 m/s2 thigh LA, 0.047 m/s2 shank LA, 0.028 deg/s thigh AV and 0.024 deg/s shank AV. All predicted trajectories were highly correlated with the measured trajectories, with correlation coefficients greater than 0.98. The motion prediction model may have a wide range of applications, such as mitigating the risk of falls or balance loss and improving the human-machine interface for wearable assistive devices.