Study On Demodulation Method In FSK-SSVEP System

Although brain-computer interface technology has developed rapidly in recent decades,the development of practical systems still faces great challenges due to the poor anti-noise ability and high complexity of EEG signals.Steady State Visual Evoked Potential(SSVEP)has been widely used because of the high recognition rate and information transmission rate.Due to the limitation of hardware technology and physiological mechanism,the stimulation frequency of SSVEP response can be counted on one hand.Consequently,this paper increases the number of stimulus instructions by using Frequency Shift Keying(FSK)encoding and increasing the instruction interval to generate TFSK(Trinary Frequency-Shift Keying Encoded)encoding.In the FSK-SSVEP system,the main problems include the difficulty of collecting useful data and the high difficulty coefficient of signal demodulation.In order to solve this problem,the specific research content of this paper is as follows:Firstly,the SSVEP modulation and demodulation methods are analyzed and the experiments of offline data acquisition are designed.(1)Based on studying the characteristics of frequency modulation and FSK modulation methods,the FSK method is selected to solve the problem that the stimulation command is limited in SSVEP for multiple frequency modulation.(2)The differences between TFSK modulation and FSK modulation in the field of communication are analyzed,and the experimental paradigm is designed by combining the characteristics of EEG signals and the variables in the modulation signal(including encoding frequency,symbol stimulus time,symbol sequence length,etc.).Offline data acquisition of 10 subjects is completed through experiments.Then,the algorithm for FSK-SSVEP signal demodulation is studied.The features of Fourier transform,wavelet transform and experiential mode decomposition algorithm are compared and analyzed.Based on the features of nonlinear,non-stationary and large noise of electroacoustic signal processing(EEG),an experiential mode decomposition algorithm is designed to carry out signal decomposition and selection and complete signal preprocessing,so as to improve the signal-to-noise ratio.The demodulation algorithm of FSK is summarized in the field of communication,and the simulation that the principles of coherent demodulation algorithm,incoherent demodulation algorithm and canonical correlation analysis algorithm are carried out.Furthermore,TFSK-SSVEP EEG data is used to analyze the influence of phase difference and starting point judgment on demodulation algorithm and demodulation results.The demodulation results show that the combined demodulation and incoherent demodulation algorithms have the best demodulation effect,and the average accuracy is as high as 93.35%.Finally,the empirical mode decomposition and incoherent demodulation algorithm are implemented on Android system,and the dialing software based on Android system is designed to complete the real-time transmission and processing of EEG signals and the algorithm is tested online.The results of successful dialing by dialing software indicate that the empirical mode decomposition and incoherent demodulation algorithms are feasible in the TFSK-SSVEP BCI system.