Research of feature extraction algorithm based on steady-state visual evoked potential

This paper introduces the EEG signals that generated by the steady-state visual evoked at brain computer interface, prompt and accurate extraction of the SSVEP signals is the key to applied it to the BCI and clinical examination, so this paper analyzes the principles and characteristics of three kinds of SSVEP feature extraction method based on fast Fourier transform, AR model power spectrum estimation and wavelet packet analysis, and designs experiments to process the EEG signals in MATLAB which generated by the stimulation of a frequency of 20HZ LED light-emitting module and extracted at the O1 channel. Through compared the processing and precision of the three methods provided a theoretical basis of reasonable choice for practical application of brain-computer interface.