Research On Key Techniques Of Hand Exoskeleton Based On Surface EMG Signal

In daily production life,diseases or accidents often cause people to lose some of their motor functions.On the other hand,population aging is becoming more and more serious,which will also have a certain impact on the normal movement of the human hand.Hand function dyskinesia and aging problems will directly cause insufficient hand movement,which seriously affect daily life.With the continuous development of robotics,robots also play an important role in medical treatment.Especially the hand exoskeleton provides a convenient way for rehabilitation of hand movement function,however,the existing hand exoskeleton devices are mainly aimed at hand rehabilitation training,while they are less applied in assisting and strengthening hand function.Therefore,it is undoubtedly of great practical significance to study a wearable mechanical device that can assist the hand to complete the daily life function to extend the rehabilitation treatment to the daily life of the patient.Furthermore,judging the action mode of the human hand by collecting the residual sEMG(surface electromyography)signal,and then controlling the movement of the exoskeleton hand to enhance the force and labor capacity,it also has a wider practical application value.The key technology of hand exoskeleton based on surface electromyography signal is studied in this thesis.The main contents include:(1)Hand exoskeleton structure design and kinematics model.The overall structure of the exoskeleton is designed based on the biological characteristics of healthy hands and referring to the size and range of motion of normal hands,including the design of the joint mechanism,the configuration of the degree of freedom,the determination of driving scheme and coupling scheme,etc.Establishing a human kinematics model and determining the relationship between the fingertip pose and the joint angles;By establishing exoskeleton kinematics and comparing the movement space of exoskeleton and human hand to verify the feasibility of exoskeleton mechanism and to make the exoskeleton movement more in line with the movement rules of human.(2)Hand motion pattern recognition based on sEMG signal.Studying and analyzing functional muscles that control different movements of human hands to determine the recognition action mode and electrode patch position according to human physiology and anatomy knowledge;Building an experimental platform to use sensors for sEMG acquisition.The original sEMG signal is preprocessed and the wavelet packet transform is used to extract the eigenvalues in the time and frequency domain.The BP neural network is constructed to recognize the motions of different motion modes.(3)Construction of hand exoskeleton motion state space under different action modes.First,The hand movement angle measuring device is designed according to the principle of human bionics.Then the physical model is constructed to measure the angular motion of each joint under different motion modes.Finally establishing the hand exoskeleton movement state space under different action modes.(4)Exoskeleton co-simulation based on Matlab and Adams.The sEMG signal acquisition system and motion recognition control system are built by Malab/Simulink to construct an exoskeleton simulation system.Controlling the hand exoskeleton model for corresponding motion based on sEMG signal.Through many simulation experiments,accuracy rate is counted by comparing the relationship curves between exoskeleton joint angle after pattern recognition and expected output angle.