Research On Intelligent Active And Passive Training Control For Upper Limb Rehabilitation

Upper limb rehabilitation has always been an important part of the rehabilitation field.With the continuous expansion of the fields of physiology and biology and the continuous improvement of the level of rehabilitation,rehabilitation training has no longer a single training method,the rehabilitation which based on the physical state of different patients provides the most appropriate rehabilitation training and rehabilitation program for the degree of illness and patient and it can automatically adjust the intensity of rehabilitation training.In this paper,the upper limb rehabilitation intelligent control system scheme is established through the analysis of the upper limb rehabilitation system,and the active and passive controllers are designed according to the needs analysis of the patients and the rehabilitation environment.The upper limb rehabilitation system was established through the analysis of the rehabilitation needs of upper limbs.Based on the dynamics and Lagrangian method,the joint dynamics model of upper limbs was established,and the model parameters was solved respectively.Based on Simulink,the upper limb human-machine joint model was established.On the other hand,the different modes of rehabilitation training was analyzed,and the upper limb rehabilitation intention identification scheme of the upper limb rehabilitation system was formulated to determine the corresponding active and passive control scheme.On the basis of biomechanics,the human-machine interaction torque characteristics and motor working characteristics corresponding was analyzed.Combined with the above characteristics and rehabilitation training evaluation indicators,the intention identification method of threshold judgment was designed.The defects of the method was analyzed.On the basis of the deficiency,the intelligent identification scheme based on neural network for active and passive intentions was established.The simulation analysis was carried out in combination with the human-machine joint model to verify the accuracy of active-passive intent recognition.According to the existing research status,the controller of the rehabilitation device was designed.The PID controller for passive and active training was designed.The requirements of the advanced controller was proposed by analyzing the actual system.An auto-disturbance controller for the rehabilitation system was designed according to training requires.The performance of the controller was simulated and analyzed.Then the controller switching algorithm was designed and simulated to verify the stability of the active and passive switching process and difficulty of controller switching.Based on d SPACE,the experimental environment of upper limb rehabilitation device was built,the calibration experiment of force sensor was set up,the characteristic experiment of active and passive controller was designed to verify simulation results.The passive constant speed rehabilitation experiment,active resistance rehabilitation experiment and neural network intention identification was designed.The experiments was designed to verify the working characteristics in the simulation,and show the effectiveness of the neural network based intent recognition scheme.