EEG Signal Processing And Its Application In BCI And Identification

Electroencephalography(EEG)signal processing technology can not only be used in clinical diagnosis and other medical research,but also can be applied in people's daily life.BrainComputer Interface(BCI)is designed to establish an output pathway between brain and computer independent of the peripheral nervous system and muscle tissue,which provides a communicate between the brain and the outside directly.BCI plays an important role in rehabilitation training of physical disability,intel igent operation,military and entertainment.On the other hand,the identification based on EEG in the field of information security has greatly enriched the biometrics with its advantages of non-replicable and non-compulsive,and has become an important complement to the modern biometrics.This paper focuses on the Steady State Visual Evoked Potential(SSVEP)based on Canonical Correlation Analysis(CCA)with portable EEG acquisition equipment and the EEG identification technology based on Empirical Mode Decomposition(EMD).SSVEP has been a research highlight in BCI due to its unique non-training,high information transfer rates and low bit error rates.CCA is one of the main methods in SSVEP because of its good performance in the feature extraction and classification.The popular EEG acquisition equipment in the laboratory are usual y expensive,bulky and complicated for using and cleaning,which hinders its further wide practical applications.Emotiv EPOC+ is a wireless EEG acquisition device with miniaturization,portability,usability,low price and so on,and is very suitable for BCI practical application research.In this paper,the CCA algorithm is studied in depth and is used to analyze the performance of SSVEP with DAQ by Emotiv EPOC+,four evoked potentials are simulated to classification.The experiment achieves 97% accuracy and information transfer rate reaches 21 bit / min.At the same time,the influence of the number of electrodes and the length of sampling time on SSVEP are analyzed.Hilbert-Huang Transform(HHT)shows its superiority in the signal processing of timefrequency analysis.The adaptive problem of Fourier transform window function's selection and wavelet basis' s selection is solved by Empirical Mode Decomposition(EMD).In this paper,EEG signal processing method based on EMD is studied for the removal of high frequency noise.This paper presents an EEG identification method based on EMD denoising.This method is proved on the BCI motion imagination competition data,and a good experimental result is obtained,especially in the individual recognition on several imagination actions.The highest recognition rate is 92.222%,which ful y embodies the feasibility and superiority of EMD in EEG identification.Finally,the influence of the recognition accuracy of different actions and the influence of the selection of electrodes on the recognition accuracy are analyzed concretely.There is a huge obstacle to the use of smart devices for people with physical consciousness and physical consciousness.Based on the above research,this paper developed a mobile phone dial-up software system based on SSVEP,which can obtain the contact telephone number through SSVEP and send it to the mobile phone to accomplish the complete dialing function.At the same time,the phone book dialing function is increased to facilitate the user finds contacts quickly through the gyroscope sensor.The software system has further standardized the EEG acquisition procedure based on Emotiv EPOC+,and can be applied to the development of other BCIs and EEG identification.