Study On Model Identification And Admittance Control Of Two Degrees Of Freedom Lower Extremity Exoskeleton

As a comprehensive product of science and technology,exoskeleton has broad application prospects in military,medical,entertainment and other fields.Due to the particularity of the exoskeleton's structure,the precise dynamic model cannot be obtained directly.Therefore,it is of great significance to study the identification of exoskeleton dynamic model and human-robot interaction control.In this paper,the lab-designed lower limb exoskeleton with two degrees of freedom is taken as the research object.Firstly,according to the limitations of traditional model-free control,the process of establishing the Lagrange dynamic model of the exoskeleton and the process of identifying the unknown parameters in the dynamic model by using intelligent swarm optimization algorithm are presented.Then,a disturbance observer based sliding mode control scheme is proposed for unknown disturbances(such as model identification error,unmodeled friction,etc.)in the actual control.Finally,three rehabilitation training modes based on admittance theory are proposed to solve the problem of man-machine interaction in actual rehabilitation medical treatment.The main contents of this paper are as follows,(1)The mechanical characteristics and working principle of the exoskeleton were analyzed,and the dynamics of the exoskeleton with two degrees of freedom was modeled by using Lagrange modeling method.Then,combining with the structural characteristics of the model,the dynamic model is linearly decoupled so that the state variables and unknown physical parameters are concentrated into the regression matrix and parameter matrix respectively.(2)The identification of dynamic parameters is discussed.Two kinds of optimization algorithms are introduced,and their principles,processes,advantages and disadvantages are explored and verified through numerical simulation.Then it discusses how to use the group optimization algorithm to identify the dynamic model,how to select the excitation trajectory,fitness function,evaluation index and so on.Finally,the platform is driven to obtain the excitation experimental data,and the lumped unknown parameters in the dynamic model are identified by the optimization algorithm and the reasonable identification results are obtained.(3)Aiming at the problems of model identification error and unknown disturbance,a disturbance observer based sliding mode control scheme was proposed.Firstly,the basic theory of sliding mode control and disturbance observer is introduced,then a nominal model based sliding mode controller is given and its stability is proved,and a nonlinear disturbance observer is introduced to observe the unknown disturbance in sliding mode control.(4)According to the specific application of exoskeleton in the field of rehabilitation medicine,the core issue of exoskeleton control is how to design targeted human-robot interaction.The differences and relations between impedance control and admittance control are explained.The influences of inertia,damping and stiffness on human-robot interaction were investigated by numerical simulation.The difference between active mode and passive mode is analyzed with admittance control theory.Finally,combining active control mode and passive control mode,a simple and effective active and passive switching strategy is proposed to help patients choose training mode independently in the middle stage of rehabilitation.