Rresearch On Motion Pattern Recognition Of Brain-controlled Upper Limb Rehabilitation Robot Based On SSVEP

Nowadays,the incidence of stroke and amyotrophic lateral sclerosis is rising,which is highly likely to lead to limb movement dysfunction after cure.Rehabilitation medical research shows that rehabilitation training and exercise can effectively assist patients to recover limb movement function.Traditional rehabilitation training exercise patients passively participate with the help of rehabilitation physiotherapists,and rehabilitation therapist industry still has a huge gap,through intelligent robot assisted patients with rehabilitation training sports has become a new effective means of rehabilitation.This paper studies a method for recognizing the motion pattern of an upper limb rehabilitation robot based on EEG signals.By identifying the EEG signals corresponding to the patient’s desired rehabilitation motion pattern,the robot is then controlled to assist the patient in performing rehabilitation training exercises.,which enable patients to actively participate in the rehabilitation training process,improving the patient’s enthusiasm and reducing the workload of the rehabilitation physiotherapist.The main work of this paper includes:(1)Analyzed the controlled upper limb rehabilitation robot,proposed the expected requirements of motion pattern recognition scheme based on rehabilitation training process,and selected steady-state visual evoked potential(SSVEP)in occipital lobe of cerebral cortex as the EEG signal for classification recognition.Based on LCD liquid crystal display,the stuttering stimulation paradigm of five targets was designed to collect SSVEP signals generated when subjects looking the video flashing targets during experiment.moreover,in order to obtain more cleaner original signal pre-processed with removed noise and power frequency was designed.(2)Canonical Correlation Analysis(CCA),multiple-variable Synchronization Index,MSI)and Filter Back Canonical Correlation Analysis(FBCCA)were used to classify the pre-processed SSVEP signals.An offline experiment was designed to collect EEG signals generated in the off-line experiment of eight subjects.Offline data were used to analyze the recognition accuracy,combination number of electrode channels and algorithm time consumption of the three algorithms under the different time window lengths.Considering the accuracy and real-time performance of the brain-control system,MSI algorithm was selected as the EEG recognition algorithm for the online experiment.(3)The 4-DOF upper limb rehabilitation robot was used for online control experiment verification,The motion pattern recognition accuracy rate of 6 subjects was up to 98%,and the information transmission rate was up to 27.33bit/min.The experimental results show that the motion pattern recognition scheme of brain-controlled upper limb rehabilitation robot based on SSVEP can achieve all the expected requirements.