Brain Computer Interface (BCI) is a system that allows users to only by brain activity to communicate with the external environment, without the use of muscle output channels. Over the years, BCI theories and methods have been developed. Currently, BCI has become a hot research projects, but there are a lot of goals to be achieved in order to improve the performance of conventional brain-computer interface system. In this paper, based on mathematical models and algorithms of P300 brain-computer interface, the main aim is to find and define best embodies the characteristics of the number P300 index (feature extraction), and then build the appropriate classification algorithm, making P300 EEG recognition accuracy as high as possible. The thesis is divided into six chapters:The first chapter describes the relevant background knowledge, including BCI connotation Introduction, significance, application technology. Furthermore, the potential for P300 and articles have been described as data instructions.The second chapter to explore data processing algorithms, and applied a set of test data to test the algorithm results.The third chapter describes several simple classification algorithm, lay the foundation for later discover better classification.Chapter IV of the main text classification method-support vector machine classification described in detail, and test the classification results.Chapter V describes Wavelet transform for signal data filtering, and test the corresponding classification results.Finally, Concludes the paper work and Looking to the future development. |