Research on Classification of Motor Imagination EEG Signals Based on Multi-domain Features

Abstract

In view of the non-stationary and nonlinear characteristics of motor imagination EEG signals, single Angle features cannot effectively reflect the comprehensiveness of signal characteristics. In this paper, a method combining time-domain features, nonlinear dynamics and spatial features is proposed. First, the original EEG signals were decomposed by variational mode (VMD), then all the modal components (IMF) were decomposed by Hilbert transform, and the average instantaneous energy of each mode was calculated as a set of features; The EEG data of each experiment were segmented to obtain the sample entropy value as a group of features. Then the common spatial pattern (CSP) processing of the EEG signals is carried out, and the CSP projection matrix is obtained through the spatial filter, and the root mean square of this matrix is taken as the third set of features. Finally, support vector machine (SVM) was used to classify EEG signals. The experimental results show that the accuracy rate reaches 94.28%, which is greatly improved compared with other methods and has a good effect.