Research On The Control Strategy Of Lower Limb Exoskeleton Movement Of Rehabilitation Robot

In recent years,as China has entered an aging society,the number of patients with spinal cord injury(SCI)has been increasing year by year,which is characterized by lower extremity motor dysfunction,inability to stand or even difficulty in self-care.Although most lower limb injuries do not cause life-threatening,they need long-term rehabilitation.Rehabilitation robot has the characteristics of objectivity,no fatigue and reliable training intensity,which has become the necessary auxiliary means of rehabilitation treatment.Therefore,the research on the lower limb exoskeleton of rehabilitation robot is of great practical significance to the patients with lower limb sports injury.Based on the provincial major project,this paper studies the control strategy of the lower limb exoskeleton movement of the rehabilitation robot.The main research work is as follows:(1)The mathematical model of the lower limb exoskeleton system of the rehabilitation robot is established.Based on the analysis of the structure and movement characteristics of the human lower limbs,the D-H method was used to obtain the spatial pose parameters of each joint of the lower limb exoskeleton system,and the kinematic analysis of the lower limb exoskeleton system of the rehabilitation robot was carried out.The Lagrangian method is used to analyze the dynamics of the lower limb exoskeleton system of the rehabilitation robot,and the mathematical model of the system is established.The curve relationship between the driving torque of each joint and the motion period is obtained,which lays the foundation for the further study of the control strategy of the lower limb exoskeleton of the rehabilitation robot.(2)The control strategy of lower limb exoskeleton movement of rehabilitation robot.As a traditional control algorithm for the exoskeleton of the lower limbs,the computational torque method is widely used in the robot field,but its control effect depends on the accuracy of the system model.Because the precise system model can’t be obtained in practical application,the control effect of the method of calculating moment is general in practical application.In order to solve this problem,iterative learning control algorithm which does not depend on the precise model of the system is selected.The closed-loop PD type iterative learning law of the lower extremity exoskeleton system is designed to prove its convergence and give the convergence conditions.The simulation results show that with the increase of the number of iterations,the trajectory tracking error of the joint angle of the lower extremity exoskeleton system under the closed-loop PD type iterative learning control is smaller than that under the calculated torque control,which shows that the iterative learning control effect is better than the calculated torque control effect It can track the desired trajectory with high precision.(3)The control strategy of the lower limb exoskeleton of the rehabilitation robot based on the initial state learning.In fact,the output of the lower extremity exoskeleton system can’t track the expected trajectory because of the fact that it can’t guarantee the same initial state as the expected trajectory,and the traditional iterative learning control algorithm can’t eliminate the influence of the initial state deviation.To solve this problem,an iterative learning control algorithm with initial state learning is adopted.On the basis of the original closed-loop PD type iterative learning law,the learning of initial state error is increased.The closed-loop PD type iterative learning law with initial state learning is designed.The convergence analysis is given based on the operator theory.The simulation results show that with the increase of the number of iterations,the deviation between the initial value of the system output curve and the initial value of the target trajectory is smaller and smaller,which shows that the improved algorithm can correct the initial error,relax the strictness of the initial conditions,and then achieve the desired motion Full tracking of the trajectory.However,the convergence rate of the system is reduced.Aiming at this problem,an iterative learning law of exponential variable gain PD type with initial learning is designed.The simulation results show that the initial convergence rate of the iterative learning control of exponential variable gain is faster than that of the iterative learning control of PD type,which can improve the convergence rate of the system tracking.