Research On Human-machine Interface Technology Of Upper Limb Rehabilitation Robot Based On SEMG And EEG

In the field of rehabilitation medical,rehabilitation robots are gradually replacing rehabilitation physicians for training and becoming the mainstream of rehabilitation training.The rehabilitation robots is becoming more and more valuable in the field of stroke rehabilitation.Human-computer interaction technology reflects the intention of the body's activities by collecting human physiological signals and converts it to the machine's control commands.The application of human-computer interaction technology in the training of rehabilitation can help patients improve the enthusiasm of training and improve the effect of rehabilitation.sEMG and EEG are the bioelectrical signals which are most commonly used in upper limb rehabilitation robots.sEMG and EEG contain a wealth of physiological information.The purpose of this paper is to study the human-computer interaction technology in upper limb rehabilitation robot based on sEMG and EEG.The study includes the pattern recognition of the 4 kinds of the movement of upper limbs and pattern recognition of EEG signals by imaging pushing and pulling.A PC system is developed.The main content of the research are:(1)Mirror therapy is used for training.sEMG signals of elbow movements is collected from contralateral upper limbs.The signal is filtered through a 5th-order band-pass filter.A valid action is detected by threshold method.Time domain feature of the motion is extracted and LDA algorithm is used to reduce the dimension of the feature matrix.The movement is identified by BP neural network with dynamic and adaptive learning rate.Experiments show that the motion achieves an average recognition rate of 99.8% by the improved BP algorithm.(2)Under the animation stimulus,we gather user's EEG signal of imaging pushing and pulling and extract the average band power of alpha and beta.The average band power of each motion is taken as the characteristic of EEG imagination.The motion achieves the highest recognition rate of 80% by BP Neural Network.(3)A upper limb rehabilitation system based on MFC framework is developed.The system can achieve online recognition of the movement of upper limb and motion instruction sending.The online recognition rate is up to 95.9%.The system verified the feasibility of this topic.Signal collecting,training and matching degree test of motion imagination are completed.The program can output the energy of the signaland the match degree between the imagined action and the training data.When the matching degree meets the requirement,the control command will be sent to the robot via the serial port.The aim of the subject is to research human-machine interface technology for stroke rehabilitation of upper limb based on sEMG and EEG.Online recognition of the movement of upper limb and online training and matching of motion imagination is realized.Mirror movement of upper limb rehabilitation training is used and the online recognition rate is 95.9% which has met the basic needs of the rehabilitation training of upper limb;The motor imagery for upper limb rehabilitation benefits the rebuilding of brain function for stroke patients.