Analysis And Research Of Brain-Computer Interface System Of Left-Right Hand Motor Imagery

Brain Computer Interface (BCI) is to establish direct communication between thehuman brain and a computer or other electronic equipment control channel. As a novelman-machine interface, BCI allows paralyzed patients to see a new hope for interactingwith the outside world independently. Because of the advantages of low cost, no injureand other feature, the noninvasive BCI has become the main research direction of thebrain-computer interface. As some research showed, BCI has a potential clinicalapplication of background on the recovery of patients with movement disorders trainingthat develop a real-time brain-machine interface on-line system has great significance.In this paper, domestic and international research background and current status ofEEG as an entry point to the left and right hand movement imagination, theevent-related desynchronization/synchronization phenomenon (ERD/ERS) offline ofthe brain motor cortex is analyzed to seek the suitable algorithms of feature extractionand pattern recognition, and then the brain-machine interface online system based onthe left and right hand movement imagination is designed and implemented.This paper mainly studies how to build up the brain-computer interface onlinesystem. The feature extraction and classification are the key technical parts ofrealization. Active One physiological signal measurement system is combined with apersonal computer to realize the BCI online system. And the experiments usingnon-invasive scalp electrodes acquisition of motor imagery EEG has good applicationprospects.In the process of experiment, subjects watch the visual stimulus which will playthe movement image of left and right hand. The left and right hand images represent themovement of different limbs imagination stimulation. At the same time, motor imageryEEG signal will be collected by Active One system, and then the signal is analyzedthrough feature extraction and classification recognition to determine which limb thesubject imagines. Estimation based on a sliding window time-domain second-ordermotor imagery EEG feature is extracted and analyzed from the perspective of energy,combined with the linear classifier based on Fisher criterion function to reduce thedimensionality of the extracted features and recognition. The experimental results showthat the brain-computer interface system can effectively extract the left and right handmovement imagination characteristics and classification and identification, determine the subjects to imagine limb, and then return the results to the man-machine interface.