Application of wavelet packet technique in BCI

Abstract

The research designed a brain-computer interface (BCI) system based on the steady state VEP potential (SSVEP). It used the wavelet packet decomposition technology to extract feature. Comparing with the FFT, this method could avoid the spectrum leakage and improve the information transmission rate. A SSVEP-based controlling system of multi-DoF manipulator was presented in this paper on the basis of virtual instruments. Various experiments showed the average detection accuracy over all subjects is about 78.75%, it is higher than the system adopted the FFT. The realization of the system provided a new exploration to the using of robot-assisted in space based on BCI.