The Research Of Asynchronous Brain-computer Interface Based On Multi-motor Imaginary EEG

Brain-computer interface is a direct communication channel which is established between the cerebrum and external devices, and is a communication control system that is not dependent on normal output channels made up of cerebral ganglion and muscle. EEG-based BCI technology is a discipline which is formed by brain science, cognitive science, rehabilitation engineering, modern information science and computer science.It achieves an exchange between the human brain information and computer or other electronic equipments. It can provide a new way of communication and control for paralysis patients, especially who lost the basic physical movements but thinking. So, it is being received increasing attention, and it also has more important scientific significance and academic value.EEG-based Imagine movement is a widely used way of thinking operations which does not require structured physical environment. EEG recognition accuracy and recognition speed of the motor imagery is the key to successful application of Brain computer interface technology. This paper summarized the basis of previous work, researched the acquisition and processing of the EEG signals.First, design the signal acquisition module for system. The paper designs imaginary-movement-task experiments. It compiles EEG acquisition software by using VC ++6.0.Second, pretreat EEG acquisition. As the EEG signal is very weak, it must be carried out on EEG denoising in the collection process. For example, we used the wavelet transform, digital filtering method to collecte signals.Third, EEG feature extraction. The purpose of feature extraction is to transform the obtained pretreatment EEG signal into different feature vector. Weather can we extract effective information in the brain activity of thinking characteristic,is the key to BCI research and the key to identify the correct sense of the basis for the different ways of thinking. This paper discusses and compares the Adaptive Autoregressive (AAR) model, Independent Component Analysis(ICA), Common Spatial Pattern(CSP) and several typical feature extraction method.Fourth, EEG classification. Classification for EEG designs is directly affecting the performance of BCI system. This article probes some classification methods, such as the Fisher linear discriminance, Neural networks, Support Vector Machine(SVM), and so on.