Research On Adaptively Interactive Control Of Lower Limb Rehabilitation Robot Based On Muscle Force Prediction

With the population aging and the increasing social pressure in China,the number of people with hemiplegia and disability caused by cardio cerebrovascular diseases is increasing day by day.Relevant rehabilitation studies have shown that the sooner such hemiplegia patients have rehabilitation training,the better their limb movement ability can be improved or even restored.On the one hand,the number of patients with hemiplegic disabilities is increasing,on the other hand,the lack of rehabilitation doctors and facilities in China is extremely scarce.Therefore,rehabilitation robot technology has gradually become a research hotspot to alleviate the lack of rehabilitation medical resources and improve the efficiency of rehabilitation training.Rehabilitation robots can complete repeated and arduous rehabilitation assisted training,reduce the labor intensity of rehabilitation physicians and improve patients' rehabilitation efficiency.But the existing rehabilitation robots are mainly used to provide passive rehabilitation for patients,and lack interaction with patients and poor rehabilitation effect and efficiency,which leads to low enthusiasm of patients to participate in rehabilitation training actively.Therefore,research on how to improve the effects of patient's active exercise intention in the lower limb rehabilitation robot control system,and how to realize the interactive control between the lower limb rehabilitation robot and patients has become an important research direction in the field of rehabilitation research.The main research work of this paper is as follows:Firstly,the human-machine integration model is established based on the analysis of the characteristics of human lower limb joint movement and lower limb rehabilitation robot structure.Then the model is reasonably simplified without affecting the law of lower limb movement.The geometric method and Lagrangian approach are respectively used to analyse kinematics and dynamics of the man-machine system,which lays the foundation for the formulation of control strategy of lower limb rehabilitation robot.Secondly,aiming at the problem of human lower limb active motion intent recognition,the method of predicting muscle force by sEMG was studied after theanalysis of the correlation coupling between sEMG and muscle force of lower limbs.Moreover,in order to solve the problems of data over-fitting and parameter optimization in the prediction model,an optimization algorithm combining cross-validation and genetic algorithm is used to optimize the model,which can improve the accuracy of prediction model in recognition of human lower limb motion intention.Thirdly,in order to realize human-machine interaction in the control system of lower limb rehabilitation robot,an interactive adaptive control strategy of lower limb rehabilitation robot based on muscle force prediction is proposed.The variable admittance controller based on muscle activity is designed to realize the adaptive tracking of human-machine interaction force.The adaptive sliding mode controller based on RBF network is designed to correct the joint angle and track the target position.And then,the interactive adaptive control method of lower extremity rehabilitation robot is verified by Matlab/simulink simulation experiment.Finally,the experimental platform of horizontal lower limb rehabilitation robot is built,witch includes signal acquisition,signal processing and rehabilitation robot control system.Then the experimental scheme and flow chart are formulated based on the proposed control strategy,and the interactive adaptive control strategy of the lower limb rehabilitation robot is applied to the experiment.The feasibility and effectiveness of the control method proposed in this paper are verified by the experiment.