Research On Algorithm Of Signal Processing In Brain-Computer Interface

Over the past decade, there has been great interest and a rapid development in the research of brain-computer interface teconology. BCI can provide a new communication option that does not depend on peripheral nerves and muscles for those with neuromuscular impairments, and might also give those without disabilities a supplementary control channel or a control channel useful in special circumstances. Signal analysis and processing is a key issue for implementation of a BCI system as well as improving its reliability and performance, and which is the main contents in the paper.As a preparative study, this paper is focused on the theory and algorithm of signal processing in BCI system. Firstly, through large numbers of reading, the state of art is reviewed, and the characters of EEG and kinds of method are summarized and compared. Then data sets provided by the BCI competitions are analyzed, and features of ERP and the applicability and ability of the methods are understood better. When taking part in the BCI competition III, T-weighted approach for feature extraction and reinforcement learning of classifier design are proposed. Compared to other methods, T-weight approach has the advantages of requiring less a prior knowledge, exploring more information and computing faster. Reinforcement learning is an optimization method both model driven and data driven aiming at mining the discriminative information as more as possible, and improving both the fitting and generalization ability of an existing classifier. Applied to the data processing, these two methods got satisfying results, and by which we won one 2nd place and three 3rd places in BCI competition III. In this paper, the mehods and results are described in detail. Finally, directions and targets in our further work about BCI especially the signal processing part are suggested.