Research On The EEG And Eye Movement Cooperative Control Of Intelligent Upper Limbs

Brain-computer interface(BCI)is a system,which bypass the consisting of muscles and nerves,and establishes a communication bridge between the brain and other devices.The system can read the intentions of brain,achieved the control with artificial limb.BCI can help those people who with motor dysfunction still have normal brain activity rehabilitation training,accomplished the control with outside devices.Electroencephalograph(EEG)signals can be used as an input source for controlling prosthetic limbs,to achieve the control of external devices.Because of the EEG signal is weak,that can easily disturbed by external environment,such as noise interference,so the error rate would be produced.The eye movement signal as a kind of biological electrical signals,produced by eye movement.Which are easy to identify and handle.By attempting found,the error rate reduced effectively by mixed eye movement signals.To solve the problem of low recognition rate of the BCI,this study presents a novel hybrid interface based on both EEG and eye movement.The proposed method provides a more effective way of identification and a wider range of applications for those patients whose muscle was damaged to communicate.The thesis includes three parts: the first part is to analyze the feature extraction of EEG,the second part is to process the eye movement signals,the last part is to analysis the pattern recognition on brain electrical signals and eye movement signals.In this thesis,for the two kinds of EEG signals based on the motor imagery,event related synchronization/ desynchronization(ERS/ERD)were tested and analyzed.The energy distribution of EEG signals tend to disperse,strongly influenced by external disturbances.The method of wavelet threshold de-noising adopted in the preprocessing part.There is an obvious advantage compared to Fourier transform.It is adaptive when dealing with digital signals,and can deal with the problem between time and frequency resolution effectively,to get the optimal time and frequency resolution.Hilbert method is used to extract the feature of the brain signals.HHT had an evident advantage compared to short time fourier transform(STFT),It can automatically extract the frequency of interest,and more accurate EEG signals can be acquired.Three kinds of common classifier algorithms have been compared,according to the compare results show that support vector machine has the highest classification and recognition accuracy.Support vector machine is selected as the classification and recognition method of EEG signals.Kalman filtering used for the eye movement signals in this thesis,gaze action were extracted by discriminant algorithm to recognize and detect fixation point.At last,the experimental results show that combination of EEG and eye movement signal can reduce the error rate of the system.A novel hybrid interface based on EEG and eye movement signals is designed in this thesis.Under the assistance of eye movement to carry out start and stop of the system has mainly studied,fixation point considered as start and stop signal of control training,when fixation point is identified,EEG command signal would interrupted.So while the customer is doing something wrong or feeling tired,the control system can be suspended through an intentional gaze action.Due to the strong independence of the two kinds of signals,through this combination,the performance of the system has improved and ensured the security of the system.