Research On Motion Prediction And Virtual Torque Control For Electro-hydraulic Driven Lower Limb Exoskeleton

Lower limb exoskeleton robot is a device that combines bionic technology and robot technology.Today,such robots are widely used in industrial production and medical rehabilitation.However,because of the complex nonlinear model of the robot and the dynamic coupling between the joints,it is difficult to realize independent control of each joint,and there is interaction interference between human and robot,both of which seriously affect the trajectory tracking accuracy of the exoskeleton.Therefore,this paper takes the lower limb exoskeleton as the research object,aiming at improving the tracking performance of the exoskeleton.The modal decoupling control is used for eliminating the dynamic coupling between joints,and the virtual torque control is used for reducing the interaction force between man-robot.The main research contents of this paper is as follows:(1)The lower limb exoskeleton kinematics model is established based on geometric relations.The dynamic model of lower limb exoskeleton is established according to Lagrange method.The hydraulic power element model is established based on the electrohydraulic servo theory.Finally,the overall model of lower limb exoskeleton system is established,which laid a foundation for the study of modal decoupling.(2)The reasons for dynamic coupling of each joint channel are analyzed,the modal space of lower limb exoskeleton system is constructed based on vibration theory,and the modal transformation matrix is obtained.The independent characteristics of each modal channel in lower extremity exoskeleton are analyzed.The physical space decoupling condition and the performance relationship between modal space and physical space are derived.Finally,the modal decoupling PI controller is designed.(3)Virtual torque control is studied to suppress the influence of human-robot interaction torque.Using Recursive Least Square method to identify the nominal model of exoskeleton.The nominal inverse model of exoskeleton is obtained by using zero-phase tracking technology.Lower limb exoskeleton motion prediction based on Kalman filter algorithm.In order to realize the design of the filter and make the robustness of the structure close to the optimal,the filter is optimized by using mixed sensitivity control.(4)Make an experimental plan for electro-hydraulic driven lower limb exoskeleton based on rapid prototyping technology.Verify the effect of modal decoupling control on electro-hydraulic driven lower limb exoskeleton.Based on modal controller to verify the virtual torque control to eliminate the effectiveness of man-machine function disturbance torque.