Search for: All records

Abstract This paper predicts the optimal motion for a repetitive lifting task considering muscle fatigue. The Denavit–Hartenberg (DH) representation is employed to characterize the two-dimensional (2D) digital human model with 10 degrees-of-freedom (DOFs). Two joint-based muscle fatigue models, i.e., a three-compartment controller (3CC) muscle fatigue model (validated for isometric tasks) and a four-compartment controller with augmented recovery (4CCr) muscle fatigue model (validated for dynamic tasks), are utilized to account for the fatigue effect due to the repetitive motion. The lifting problem is formulated mathematically as an optimization problem, with the objective of minimizing dynamic effort and joint acceleration subjected to both physical and task-specific constraints. The design variables include joint angle profiles, discretized by quartic B-splines, and the control points of the profiles of the fatigue compartments associated with major body joints (spinal, shoulder, elbow, hip, and knee joints). The outcomes of the simulation encompass profiles of joint angles, joint torques, and the advancement of joint fatigue. It is notable that the profiles of joint angles and torques exhibit distinct periodic patterns. Numerical simulations and experiments with a 20 kg box reveal that the maximum predicted lifting cycles are 11 for the 3CC fatigue model and 13 for the 4CCr fatigue model while the experimental result is 13 cycles. The results indicate that the 4CCr muscle fatigue model provides enhanced accuracy over the 3CC model for predicting task duration (number of cycles) of repetitive lifting. 

IntroductionCompartment based models of muscle fatigue have been particularly successful in accurately modeling isometric (static) tasks or actions. However, dynamic actions, which make up most everyday movements, are governed by different central and peripheral processes, and must therefore be modeled in a manner accounting for the differences in the responsible mechanisms. In the literature, a three-component controller (3CC) muscle fatigue model (MFM) has been proposed and validated for static tasks. A recent study reported a four-compartment muscle fatigue model considering both short- and long-term fatigued states. However, neither has been validated for both static and dynamic tasks. MethodsIn this work we proposed a new four-compartment controller model of muscle fatigue with enhanced recovery (4CCr) that allows the modeling of central and peripheral fatigue separately and estimates strength decline for static and dynamic tasks. Joint velocity was used as an indicator of the degree of contribution of either mechanism. Model parameters were estimated from part of the experimental data and finally, the model was validated through the rest of experimental data that were not used for parameter estimation. ResultsThe 3CC model cannot capture the fatigue phenomenon that the velocity of contraction would affect isometric strength measurements as shown in experimental data. The new 4CCr model maintains the predictions of the extensively validated 3CC model for static tasks but provides divergent predictions for isokinetic activities (increasing fatigue with increasing velocity) in line with experimentally observed trends. This new 4CCr model can be extended to various domains such as individual muscle fatigue, motor units’ fatigue, and joint-based fatigue. 

Abstract This study introduces a hybrid model that utilizes a model-based optimization method to generate training data and an artificial neural network (ANN)-based learning method to offer real-time exoskeleton support in lifting activities. For the model-based optimization method, the torque of the knee exoskeleton and the optimal lifting motion are predicted utilizing a two-dimensional (2D) human–exoskeleton model. The control points for exoskeleton motor current profiles and human joint angle profiles from cubic B-spline interpolation represent the design variables. Minimizing the square of the normalized human joint torque is considered as the cost function. Subsequently, the lifting optimization problem is tackled using a sequential quadratic programming (SQP) algorithm in sparse nonlinear optimizer (SNOPT). For the learning-based approach, the learning-based control model is trained using the general regression neural network (GRNN). The anthropometric parameters of the human subjects and lifting boundary postures are used as input parameters, while the control points for exoskeleton torque are treated as output parameters. Once trained, the learning-based control model can provide exoskeleton assistive torque in real time for lifting tasks. Two test subjects’ joint angles and ground reaction forces (GRFs) comparisons are presented between the experimental and simulation results. Furthermore, the utilization of exoskeletons significantly reduces activations of the four knee extensor and flexor muscles compared to lifting without the exoskeletons for both subjects. Overall, the learning-based control method can generate assistive torque profiles in real time and faster than the model-based optimal control approach. 

In this study, we introduce a two-dimensional (2D) human skeletal model coupled with knee, spine, and shoulder exoskeletons. The primary purpose of this model is to predict the optimal lifting motion and provide torque support from the exoskeleton through the utilization of inverse dynamics optimization. The kinematics and dynamics of the human model are expressed using the Denavit–Hartenberg (DH) representation. The lifting optimization formulation integrates the electromechanical dynamics of the DC motors in the exoskeletons of the knee, spine, and shoulder. The design variables for this study include human joint angle profiles and exoskeleton motor current profiles. The optimization objective is to minimize the squared normalized human joint torques, subject to physical and task-specific lifting constraints. We solve this optimization problem using the gradient-based optimizer SNOPT. Our results include a comparison of predicted human joint angle profiles, joint torque profiles, and ground reaction force (GRF) profiles between lifting tasks with and without exoskeleton assistance. We also explore various combinations of exoskeletons for the knee, spine, and shoulder. By resolving the lifting optimization problems, we designed the optimal torques for the exoskeletons located at the knee, spine, and shoulder. It was found that the support from the exoskeletons substantially lowers the torque levels in human joints. Additionally, we conducted experiments only on the knee exoskeleton. Experimental data indicated that using the knee exoskeleton decreases the muscle activation peaks by 35.00%, 10.03%, 22.12%, 30.14%, 16.77%, and 25.71% for muscles of the erector spinae, latissimus dorsi, vastus medialis, vastus lateralis, rectus femoris, and biceps femoris, respectively. 

In this study, a novel human-in-the-loop design method using a genetic algorithm (GA) is presented to design a low-cost and easy-to-use four-bar linkage medical device for upper limb muscle rehabilitation. The four-bar linkage can generate a variety of coupler point trajectories by using different link lengths. For this medical device, patients grab the coupler point handle and rotate the arm along the designed coupler point trajectory to exercise upper limb muscles. The design procedures include three basic steps: First, for a set of link lengths, a complete coupler point trajectory is generated from four-bar linkage kinematics; second, optimization-based motion prediction is utilized to predict arm motion (joint angle profiles) subjected to hand grasping and joint angle limit constraints; third, the predicted joint angles and given hand forces are imported into an OpenSim musculoskeletal arm model to calculate the muscle forces and activations by using the OpenSim static optimization. In the GA optimization formulation, the design variables are the four-bar link lengths. The objective function is to maximize a specific muscle’s exertion for a complete arm rotation. Finally, different four-bar configurations are designed for different muscle strength exercises. The proposed human-in-the-loop design approach successfully integrates GA with linkage kinematics, arm motion prediction, and OpenSim static optimization for four-bar linkage design for upper limb muscle strength rehabilitation. 

Abstract In this study, the fatigue progression and optimal motion trajectory during repetitive lifting task is predicted by using a 10 degrees of freedom (DOFs) two-dimensional (2D) digital human model and a three-compartment controller (3CC) fatigue model. The numerical analysis is further validated by conducting an experiment under similar conditions. The human is modeled using Denavit-Hartenberg (DH) representation. The task is mathematically formulated as a nonlinear optimization problem where the dynamic effort of the joints is minimized subjected to physical and task specific constraints. A sequential quadratic programming method is used for the optimization process. The design variables include control points of (1) quartic B-splines of the joint angle profiles; and (2) the three compartment sizes profiles for the six physical joints of interest — spine, shoulder, elbow, hip, knee, and ankle. Both numerical and experimental liftings are performed with a 15.2 kg box as external load. The simulation reports the human joint torque profiles and the progression of joint fatigue. The joint torque profiles show periodic trends. A maximum of 17 cycles are predicted before the repetitive lifting task fails, which also reasonably agrees with that of the experimental results (16 cycles). This formulation is also a generalized one, hence it can be used for other repetitive motion studies as well. 

The ability to predict the decline in muscle strength over the course of an activity (i.e., fatigue) can be a crucial aid to task design, injury prevention, and rehabilitation efforts. Current models of muscle fatigue have been hitherto validated only for isometric contractions, but most real-world tasks are dynamic in nature, involving continuously varying joint velocities. It has previously been proposed that a three-compartment-controller (3CC) model might be used to predict fatigue for such tasks by using it in conjunction with joint- and direction-specific torque-velocity-angle (TVA) surfaces. This allows for the calculation of a time-varying target load parameter that can be used by the 3CC model, but it increases model complexity and has not been validated by experimental data. An alternative approach is proposed where the effect of joint velocity is modeled by a velocity parameter and integrated into the fatigue model equations, removing the dependence on external TVA surfaces. The predictions using both methods are contrasted against experimental data collected from 20 subjects in a series of isokinetic tests involving the knee and shoulder joints, covering a range of velocities encountered in day-to-day tasks. A much lower degree of fatigue is observed for moderate velocities compared to that for very low or very high velocities. Predictions using the integrated velocity parameter are computationally less expensive than using TVA surfaces and are also closer to experimentally obtained values. The modified fatigue model can therefore be applied to dynamic tasks with varying velocities when the task is discretized into several isokinetic tasks.