基于滑动窗ICA的信号包络检测及其在脑－机接口中的应用

he mental state has a close relationship to the envelop changes of basic rhythms of EEG and this feature is widely used in algorithm design and system implementation of brain-computer interface(BCI). In this paper, we put forward a new method of applying ICA algorithm based on sliding window to the online envelope extraction of signals. By a detailed analysis and demonstration of the efficiency of this algorithm, we find that ICA algorithm based on sliding window not only has a better blind source separation performance, but also can give the amplitude modulation information of the source signal by the ICA dynamic mixture model parameters. This paper applies the proposed algorithm to envelope extraction of the mu rhythm of EEG and motor imagery classification. At last, the implementation and testing of an online BCI system based on ICA algorithm is described and a better online recognition results are achieved.