The Identification Of Nonlinear And Muscle-activation-dependent Human-exoskeleton Coupling Model

Lower limb exoskeletons are mechanical devices used for human motion assistance and functional enhancement.As a typical human-exoskeleton integrated system,the coupling force between humans and exoskeletons directly affects the wearer’s comfort and the control effect of the exoskeleton.Most previous studies focused on the control of the exoskeleton,treating the human-exoskeleton coupling force as a disturbance in the human-exoskeleton integrated system,aiming to achieve better control effects,but ignoring the impact of human-exoskeleton coupling force on the coupling safety and comfort of the exoskeleton.Therefore,accurately predicting human-exoskeleton coupling force is of great significance for improving the control effect of exoskeletons and improving the comfort of wearers.An accurate human-exoskeleton coupling dynamics model is a prerequisite for accurate prediction of human-exoskeleton coupling forces.In previous studies,the humanexoskeleton coupling model was modeled as a linear spring-damped system to improve the control of the exoskeleton.However,in-depth studies have revealed the limitations of the linear model.To explore the human-exoskeleton coupling dynamics model in depth,this paper proposes a nonlinear human-exoskeleton coupling model and a muscle activationdependent human-exoskeleton coupling model on the basis of the linear model.In order to improve the applicability of the model in looser coupling systems,this thesis proposes a nonlinear human-exoskeleton coupling model and its parameter identification method.Comparing the error of predicting coupling force between linear and nonlinear models,it is proved that nonlinear models have better applicability in looser coupling states.Combined with the analysis of human-exoskeleton coupling experimental results,it reveals the correlation between coupling parameters of nonlinear models and coupling position and tightness.Based on coupling position,coupling tightness and anthropometric parameters,a regression model of human-exoskeleton coupling parameters is established.By analyzing the sensitivity of regression models to their inputs,we evaluate the importance of each input parameter to human-exoskeleton coupling parameters.In order to improve the applicability of the model in variable-parameter coupling systems,this thesis proposes a muscle-activation-dependent human-exoskeleton coupling model and its parameter identification method based on linear models.Through analysis of muscle characteristics in coupling areas,EMG signals during muscle group activation on anterior thigh,posterior thigh,anterior calf and posterior calf were reconstructed respectively with human-exoskeleton coupling experimental data to obtain muscle-activationdependent human-exoskeleton coupling parameters.By comparing muscle-activationdependent human-exoskeleton coupling predictive force with linear model’s human-exoskeleton coupling predictive force,muscle-activation-dependent model is superior.Based on muscleactivation-dependent human-exoskeleton coupling experimental results,we analyze the relationship between human-exoskeleton coupling parameters and muscle activation.The study of the human-exoskeleton coupling model based on nonlinear and muscleactivation-dependent improves the applicability of the model and the accuracy of the model in predicting human-exoskeleton coupling forces,while revealing the mechanical constitutive and operation mechanism of the human-exoskeleton coupling region.