A Blind Deconvolution Method of Single-Channel the EEG Convolution Mixed Signals Based on the Linear Predication Error Analysis

EEG signal has the chaotic character. Based on the chaotic characteristics and the linear predication error analysis, a deconvolution filter for single-input and single-output chaotic convolution mixed signal is discussed. Then corrected estimated error from the output data of this filter by the singular spectrum analysis. So the blind deconvolution of the source signals and the transmission function in the single-channel chaotic convolution system can be realized. This method takes advantage of the chaotic physical features and the compensatory technique that is based on the phase space reconstruction chaos dynamic equation. And EEG signal simulation result verified the effectiveness of this proposed method.