The Research Of EEG Recognition Based On Vibration Assistance

The Brain Computer Interface(BCI)is a direct connection between brain and external devices,and it is an effective way for brain-damaged patients to interact with the outside world.Motion imagery EEG signal is the most commonly used control signal in BCI system.External evoked stimuli are transmitted to the motor cortex of the brain through mechanical receptors on the human skin surface,which has become a research hotspot to improve the recognition rate of BCI.This paper studies whether vibration assistance can affect the activity level of brain motor cortex,and whether it can improve the recognition rate of EEG signals.The main contents and innovations are summarized as follows:(1)In terms of signal de-noising,aiming at the phenomenon of mode mixing phenomenon existing in the traditional Empirical Mode Decomposition(EMD)and the uncertainty of the sequence of independent source components obtained by the traditional Blind Source Separation,a new method of EEG de-noising based on the Ensemble Empirical Mode Decomposition(EEMD),De-noising Source Separation(DSS)and Approximate Entropy(ApEn)is proposed.The validity of the proposed method was confirmed by simulation EEG data and real experiment EEG data.(2)In the analysis of the state of the motor cortex,based on the ERD/ERS and the influence of the alpha and beta rhythm signals affected by the motion image and somatosensory stimulation,two quantified ERD values of the alpha and beta rhythm signals in the time domain and the frequency domain were used to analyze the state of motor cortex EEG signal under the aid of a single motor imagination and different vibration frequencies in order to measure the activivity of motor cortex.The experimental data of four subjects were analyzed by statistical analysis.The results showed that vibration-assisted stimuli increased the activity of motor cortex and enhanced the resolution of EEG signal.(3)In the aspect of feature extraction,aiming at the problem that the quality factor(the ratio of center frequency to bandwidth)of the wavelet basis function of traditional wavelet transform cannot be adjusted flexibly,a method of EEG feature extraction based on Wavelet Transform With Tunable Q-Factor(TQWT)and fuzzy entropy is proposed.The effectiveness of the proposed method was verified by the international BCI competition data.(4)In the aspect of pattern recognition,the Extreme Learning Machine(ELM)classification method based on Particle Swarm Optimization(PSO)was applied to optimize the input layer nodes weight and the hidden layer nodes threshold of the ELM,so that the optimized extreme learning machine could obtain better classification results only in fewer hidden nodes.(5)In the research of EEG signal recognition under vibration assistance,the EEG signals of left and right hand movement of four subjects under single motion imagery and vibration-assisted motion imagery were studied experimentally.The results showed that the average of four subjects' recognition accuracy of the left and right hands with vibration-assisted stimuli increased by 5%and 7%compared with the single motion image.