Multi-class Motor Imagery Recognition of Single Joint in Upper Limb Based on Bispectrum

Aiming at the difficulties in extracting effective features and low classification accuracy in the current multiclass motor imagery recognition, this paper proposes a multiclass motor imagery recognition method based on bispectrum analysis and twin support vector machine (TWSVM). First, the bispectrum analysis method is used to extract the features of the signal; finally, the features are sent to TWSVM for classification of the EEG, obtain an average recognition rate of 78.57%, which provides an effective method for multi-class motor imagery recognition, which will greatly pro-mote in practical application based on BCI.