Control System Design Of Lower Limb Rehabilitation Robot Based On Human Gait Characteristics

Medical and clinical trials,and therapist assisted rehabilitation training,motor function restoration have enabled relatively safe and effective for patients with lower limb movement dysfunction.However,with the aggravation of population aging,the patients have increased dramatically,which has far exceeded the carrying capacity of therapist.The emergence of lower limb rehabilitation robot(LLRR)brings hope to patients.Patients will get effective rehabilitation training and restore normal walk with the help of non-therapist using LLRR.The LLRR has combined with robotics,rehabilitation medicine and biomechanics and other fields.To ensure the safety and effectiveness of rehabilitation training for patients,it is necessary to acquire normal human gait data(HGD),and apply the data to the control system design after optimization.How to obtain detailed data of normal human gait,establish a humanoid model,optimize the tracking performance of the controller and apply the designed control strategy to rehabilitation training have became a hot spot in the research field of LLRR.Therefore,dynamic modeling of LLRR and the optimization design of humanoid control system based on HGD are discussed in this paper.The main research contents are shown as follows.(1)To realize the humanoid control system design of lower limb rehabilitation robot,the human motion mechanism is adopted to design the control system in this paper.The HGD of healthy subject is collected using three-dimensional motion capture system.The angles of human hip joint and knee joint are obtained after kinematics calculation,which are the expected human gait trajectories.The foot reaction force of the subject is collected using the three-dimensional force measuring platform.The joint torque is obtained after kinematics calculation,which is fed to the LLRR.The LLRR can move like the normal human based on the HGD and the joint torque.(2)To ensure the assisted rehabilitation training of the LLRR more consistent with the human motion characteristics,and to improve the safety and effectiveness of the rehabilitation training,three controllers based on the HGD for LLRR are proposed.The controllers are adaptive PD controller,adaptive radial basis function(RBF)neural network controller and adaptive RBF neural network controller based on feed-forward control,respectively.The adaptive PD controller takes gravity and unknown disturbance into account and ignores the friction force.The gravity,friction,and unknown disturbances have been considered in the other two controllers.The simulation results using Matlab show that all the three control methods can track the desire trajectory.As shown in the results,the tracking speed and tracking effect of the RBF neural network adaptive controller are much better than that of the PD adaptive controller,and the initial input torque is smaller.,The tracking speed and tracking effect of the RBF neural network adaptive controller have been improved after adding the feedforward controller.(3)In this paper a control scheme for LLRR based on HGD and joint torque is proposed.First,the real normal HGD is used as the reference trajectory of the control system,and the joint torque is fed to the LLRR.Then,a PD controller based on joint angle and angular velocity tracking error is designed to compensate the input torque of LLRR,which can realize the LLRR tracking the desire trajectories.Finally,the effectiveness of the proposed method is confirmed based on the simulation results using Matlab.