DWT and CNN based multi-class motor imagery electroencephalographic signal recognition

摘要

Objective. Brain computer interface (BCI) system allows humans to control external devices throughmotor imagery (MI) signals. However, many existing feature extraction algorithms cannot eliminatethe influence of individual differences. This research proposed a new processing algorithm thatcan reduce the impact of individual differences on classification and improve the universality ofthe algorithm. Approach. To select the optimal frequency band, the energy in each sub-band wascalculated by the discrete wavelet transform. Power spectral density and visual geometric groupnetwork based convolutional neural network were used for feature extraction and classificationrespectively. Main results. The test of the BCI Competition IV dataset IIa proved the superiority ofthe algorithm. In comparison with some commonly used methods, the proposed algorithm reducedclassification calculation time while improving classification accuracy; the average classificationaccuracy rate reaches 96.21%, which is far exceeding the results obtained by the latest literature.Significance. The good classification performance of this research was rooted in the reduced numberof parameters, the reduced consumption of computing resources, and the eliminated influence ofindividual differences. Therefore, the proposed algorithm can be applied to a real-time multi-classBCI system.