Abstract

Info

Author(s):: FENG Qinchang; Guangdong Institude of Medical Instruments, Guangzhou 510500, China

Keywords:: Keywords: brain-computer interface; electroencephalogram signal; OpenBCI; OpenViBE; motor imagery

Abstract:: Abstract: A motor imagery brain-computer interface for the control of mechanical arms is designed based on the μ rhythm and β rhythm from the ensorimotor cortex of brain, with OpenBCI as electroencephalogram (EEG) signal acquisition platform, OpenViBE as EEG analysis platform, common spatial pattern algorithm for feature extraction, and Gaussian kernel support vector machine for feature classification. Moreover, the performances of signal feature extraction method and classification algorithm were evaluated by experiments. The preliminary experimental results reveal that the classification accuracy after the signal extraction with common spatial pattern algorithm is higher than that after the signal extraction with surface Laplacian spatial filter, and that the classification performance of support vector machine is superior to linear discriminant analysis. The designed system can realize the effective control of mechanical arms, with a control accuracy higher than 95%. Future studies will focus on improving the hardware control functions of OpenViBE by custom plug-in.