Research On Interactive Control Method Of Lower Limb Rehabilitation Robot Based On Rigid-Flexible Coupling

With the advancement of society and the development of technology,robots have been closely related to all aspects of people’s food,clothing,housing and transportation,and at the same time have a significant influence on the medical industry.Rehabilitation robots are mainly used to help patients with movement disorders complete related rehabilitation training content,and have the ability to feed back limb information to patients and doctors.The rehabilitation robot and the patient drive the affected limb to complete the rehabilitation training of different actions in the same working space,so that the affected limb gradually recovers the lost movement function.In this paper,a wearable lower limb rehabilitation robot based on rigid-flexible coupling is designed based on the principles of biology and human anatomy,combined with the joint structure of the human lower limbs,which can assist patients in lower limb rehabilitation training.The main work is as follows:1.By analyzing the composition of human bones,combined with the joints of the lower limbs,the three joints of the lower limb rehabilitation robot are designed for the structure of the hip,knee and ankle joints,and a flexible device(shock absorber)is used to simulate the articular cartilage of the human body.Completed the structural design of the lower extremity rehabilitation robot based on CATIA software,imported the design parameters into the ADAMS software for simulation analysis,established the three-joint dynamic model of the lower extremity based on the lagrange principle,and verified the rationality of the structural design of the lower limbs rehabilitation robot through experiments;2.(Adaptive Fuzzy Neural Network,AFNN)intention recognition method is proposed.By analyzing the signal of(Surface Electromyography,s EMG),obtaining the information of the lower limb movement intention of the human body,and comparing the lower limb joint angle measured by the angle sensor,the validity of the proposed AFNN recognition method is verified.Compared with the BP and RBF neural network prediction models,the results show that the prediction accuracy of the proposed method is increased by 53.37% to 92.93%,and the convergence speed is increased by 36.66% and 80.49%.3.Aiming at the human-computer interaction control problem,combined with the established dynamic model of the lower limb rehabilitation robot,a control algorithm based on the combination of gradient descent method and neural network is designed.By comparing with PID control algorithm,the effectiveness of the proposed control algorithm is verified.Taking into account the disturbance problem of the lower limb rehabilitation robot in the rehabilitation process,the interference always exists in the rehabilitation training process,numerical experiments have proved that the proposed algorithm has better interference suppression ability,and has better stability than the PID control algorithm.