Research And Application Of Motor Imagery Brain-computer Interface Based On Speech Imagery Enhancement

Brain-computer interface(BCI)is a technology that does not rely on peripheral nerves and muscles to realize brain-machine communication.It has broad application prospects in medical rehabilitation,education,entertainment,and military fields.Among them,the motor imagery BCI is less traumatic and easy to operate.It has great potential and research value.This paper analyzes the feature extraction of EEG signals,experimental paradigm and online system,designs and implements a motor imagery BCI control system with speech imagery enhancement function.The specific research content is as follows:(1)The WPD-ACSP feature extraction algorithm is proposed.On the basis of the common spatial pattern algorithm,the wavelet packet decomposition is combined with the adaptive common spatial pattern algorithm.Firstly,the motor imagery EEG signal is decomposed by four layers of wavelet packet,and then the sub band S(4,1)related to motor imagery is selected for reconstruction to obtain the relevant frequency-domain information,and then the reconstructed signal is sent to the spatial filter for feature extraction.Experimental results show that this method can effectively extract the frequency domain and spatial domain features of motor imagery EEG signals,and improve the classification accuracy.(2)The experimental paradigm of motor imagery based on speech imagery enhancement is designed,which uses the reading and writing imagery task based on Chinese pinyin to replace the traditional right hand motor imagery.When the subjects perform this task,they will produce ERD/ERS phenomenon of speech imagery and motor imagery at the same time,and then enhance the features of EEG signals.A total of 8 subjects were recruited for off-line experiment,and the results were analyzed by ERSP,brain topographic map and classification.The experimental results show that speech imagery can effectively enhance the features of EEG signals,and improve the class differentiation between the two types of signals.The average classification accuracy of the improved paradigm is 8% higher than that of the traditional paradigm.(3)An on-line control system of brain computer interface(BCI)is built,which combines the multi joint robot arm based on ROS as the lower computer with the motor imagery BCI enhanced by speech imagery.Two kinds of control strategies are adopted to complete the experiment of on-line control of robot arm movement and grasping,which improves the stability of BCI on-line control system.