Research On Brain-computer Interface Based On Hybrid-mode

Brain-computer interface is a kind of control technology that can directly connect the brain with the external environment.The implementation process only rely on the electrical activity generated when the brain makes corresponding thinking to converted into a corresponding Control orders instead of the peripheral nerves and muscle tissue,to communicate with the outside world.EEG signals reflect the functional status of the brain as well as brain electrical activity and are the most widely used tools for studying brain-computer interface technology.Recently,the research of BCI is mainly for single-mode EEG signal.The BCI based on motor imagery ERD/ERS is the most common.As a spontaneous EEG,ERD/ERS has a strong randomness and nonstationarity,the characteristics of EEG are not obvious and the different tasks can be distinguished poorly and so on.To solve these problems of single-mode simple motor imagery BCI,this paper proposed a complex mental task model of motor imagery,a new hybrid BCI that combined motor imagery with special mental tasks,which used special mental tasks instead of simple motor imagery tasks to enhance ERD/ERS physiology.The main findings include:Firstly,the typical BCI system based on EEG signal were comparatively studied.Focusing on the limitations and pertinence of single-mode systems,the hybrid mode was explored.Based on the hybrid-mode BCI system,the ERD/ERS mode of single motor imagery was taken as the research object.In view of the fact that singleness of feature information,ERD/ERS physiological phenomenon is not obvious and poor distinguish ability,a new type of complex hybrid BCI that combines sports imagination with mental tasks was proposed.Secondly,Targeting at the EEG signal of ERD/ERS,several common feature extraction and pattern classification methods were designed,including Auto Regressive,Common Spatial Pattern filter,Bayesian classifier,Fisher linear classifier and Sparse Representation classification.Combining each feature extraction and classification method to test the competition data based on motor imagery and compared the classification results to find the more suitable method for processing the EEG signal of ERD/ERS,then this method can be used to process complex hybrid EEG signal.Lastly,The experimental design of the new complex hybrid BCI was carried out,including subjects,experimental process and EEG signal acquisition.At the same time,in order to verify the validity of the proposed new complex hybrid-model,a simple motor imagery mode was designed as a contrast experiment.EEG signals of two modes are processed using the designed feature extraction and classification methods.The experimental result show that the performance of the complex hybrid BCI based on motor imagery and special mental tasks is superior to that of BCI system based on motor imagery.