Research on Recognition of Shoulder Joint Movement Imagination Based on BCI Technology

Abstract

In the research of brain-computer interface (BCI) technology, it is difficult to recognize electroencephalogram (EEG) signals induced by different movements of a single joint. In response to this situation, this paper proposes the use of empirical mode decomposition (EMD) to obtain the intrinsic mode function (IMF), combining amplitude-frequency (AF) domain information in the IMF with the common spatial pattern (CSP) to propose the AF-CSP construction shoulder. The eigenvectors of the three types of joints imaginary EEG signals are classified and recognized by the twin support vector machine. The experimental results show that the accuracy rate of the feature vectors constructed by the AF-CSP method proposed in this paper is 89.7 percent recognized by the twin support vector machine, which proves the effectiveness of the method and can be further used in brain-computer interfaces.