SSVEP EEG signal classification model based on EEMD-FastICA-TRCA

In order to retain as many effective signals in EEG as possible and improve the accuracy of a signal classification and the performance of the brain-computer interface system, a steady-state visual evoked potential EEG classification model combining ensemble empirical mode decomposition, fast, independent component analysis and task-related component analysis was proposed to solve the problem that noise would reduce the accuracy of signal classification and recognition. Firstly, it is proved that the proposed EEMD-FastICA noise reduction model has certain advantages over the standard noise reduction methods, and the performance of the model is verified by using the Benchmark dataset of the brain-computer interface group of Tsinghua University. On this basis, the EEMD-FastICA noise reduction model is combined with the TRCA algorithm to classify EEG signals, and the classification results are compared with other classification models. The comparison results show that under different time Windows, the classification accuracy and information transmission rate of the proposed model are higher than those of other models. Within 1.5 S, the highest classification recognition accuracy was 85.83%, and the highest ITR was 238.34 bits/min.