Research On Modeling And Control Strategies Of Lower Extremity Exoskeleton Robots

Nowadays,our country is facing the acceleration of aging process,the increase of disabled and the lack of individual combat ability.In this case,the exoskeleton system was developed.The exoskeleton is a kind of equipment with bionic performance and it is the combination of artificial intelligence and robotics.They are widely used in medical rehabilitation,military,agriculture,industry and other fields.In order to help the elderly and the disabled to restore or enhance the ability to walk,a lower limb exoskeleton was proposed in this paper.It has 5 DOFs in each leg.Theoretical analysis and simulation verification were used in this paper to deal with the structure design and intelligent control system.The details are as follows:(1)Analysis of present situation and structure design of lower limb exoskeleton.The existing exoskeleton system was analyzed firstly in this paper and the advantages/disadvantages was illustrated.Then the research content was clearly.Then based on physiological structure and gait parameters of human leg,the freedom and mechanical structure were designed for the lower limb exoskeleton leg with 5 DOFs in each leg.The related parameters of lower limb exoskeleton are also given in this paper.(2)Kinematics of the lower limb exoskeleton.The direct kinematic model of left/right leg in lower limb exoskeleton was established by Craig modified version of the D-H convention and it was verified by MATLAB/Robotics toolbox.Based on the verified direct kinematic model,the inverse kinematic model of lower limb exoskeleton was established by algebraic method,and it was verified by MATLAB/Simulink.At the same time,the workspace of lower limb exoskeleton with 5 DOFs was analyzed by Monte Carlo method.Then the kinematic model can be further verified.(3)Dynamics and virtual prototyping of the lower limb exoskeleton.The dynamic mathematical model of lower limb exoskeleton was established by Lagrange equation.But the mathematical model was nonlinear and highly coupled.Thus,this paper used the object oriented to establish virtual prototype.The three-dimensional model of lower limb exoskeleton was established by Solidworks and can be imported into MATLAB/SimMechanics to build virtual prototype.Finally,the mathematical model was created by Robotics Toolbox and be compared with virtual prototype of lower extremity exoskeleton.The comparison verified that virtual prototype was valid.(3)Design of a model-free based neural network control with time-delay estimation.According to the nonlinear and higly coupled mathematical model of lower limb exoskeleton,the model-free control based on algebraic method was introduced.The ultra-local model was used as the system model,and the time-delay estimation was used for unknown parts.Then the time-delay estimation model-free controller(TDE-MFC)can be gained.Further,an adaptive RBF neural network was introduced to compensate the estimation error.Then comes to the model-free based neural network control with time-delay estimation(TDE-MFNNC).(5)Simulation of TDE-MFNNC based on virtual prototype.The time-delay was chosen through the comparison simulation.Based on the delay,the proposed method(TDE-MFNNC)was simulated and tested on virtual prototype to comparing with PD controller,neural network and TDE-MFC.Comparisons were in the ideal case and the disturbance case.This comparative study validated TDE-MFNNC as more stable and effective than the traditional controllers.Further,the kinematic model of lower limb exoskeleton added to exoskeleton also proved the stability and effectiveness of the proposed controller.