The Design And Development Of A Hybrid Brain Computer Interface Based On SSVEP And EMG

Brain-computer interface(BCI)technology is applied in a human-computer interaction system that does not depend on peripheral nerves and muscle tissue.The application of BCI can greatly improve the living ability of patients with motor dysfunction,which plays an important role in medical rehabilitation and life assistance.At present,the research of BCI mainly focuses on a single type EEG signal,resulting in some shortcomings in practical application,etc.poor reliability,simple recognition command,low accuracy and slow speed of signal data processing.A hybrid BCI system based on steady-state visual evoked potential(SSVEP)and electromyography(EMG)is proposed in the thesis.The corresponding signal processing algorithms are studied.The main research work subsequently is described as follows:(1)The system experiment is designed and completed.The SSVEP stimulation paradigm of five stimulation blocks is designed and implemented by analyzing the SSVEP stimulation methods,stimulation frequency,stimulation target size and layout.The effects of time window length,number of electrodes and position on the correct classification rate of the experiment are analyzed,then the optimal electrode collection channel is selected.The glitches signal state detector is designed to feed-back the error output commands based on the characteristics of the EMG signal.(2)The signal processing algorithms are investigated.The discrete wavelet denoising method is utilized to remove the low-frequency interference in the SSVEP signal.Then the Canonical Correlation Analysis(CCA)algorithm is applied to realize feature extraction and classification identification.To solve the problem that CCA algorithm can’t extract multi-dimensional information from signal well,CCA algorithm is improved to extracte signal features more effectively and quickly.Discrete wavelet denoising is utilized to denoise in EMG signal.Linear classifier and support vector machine(SVM)classifier are studied according to signal characteristics.The K-fold cross-validation method is adopted for classifier training and parameter selecting in the SVM classifier,which improves the performance of the SVM classifier and achieves better classification results.(3)An online hybrid BCI system is designed and constructed to realize flight control of multiple commands on the three-degree-of-freedom helicopter(3-DOF helicopter)platform.Six subjects participates in the online system experiment,and success to pick up and place objects,with an average accuracy of 90.00±4.30%.The hybrid BCI system based on SSVEP and EMG is proposed in this thesis.The online control of the 3-DOF helicopter is realized to verify the effectiveness and stability of the system.The research results will play a huge role in medical rehabilitation,life assistance and so on.