Research On Application Of EEG Recognition In Intelligent Prostheses Field

Brain-computer Interface (BCI) is a kind of information transmission channel which is not dependent on human peripheral nerves and muscles. It puts forward a new kind of method to communicate with the external environment. Brain-computer Interface based on EEG has the advantages of low cost, convenient operations, no damages to the brain etc. It is the main research direction of Brain-computer Interface. The research of Brain-computer Interface has clear application background, therefore, it has very important practical significance to study Brain-computer Interface about connections and controls of the lower robotic bionic legs.This paper respectively aims at BCI based on the steady state visual evoked potential (SSVEP), BCI based on the alpha wave block, BCI based on the motor imagery and conducts a deep study and discussion on the three directions.The independent component analysis (ICA) algorithm, wavelet analysis and fast Fourier transform (FFT) has been innovatively applicated in the steady-state visual evoked potential in this paper. At the same time, we also design the phase-frequency encoding, which results is much better then hours division coding and frequency division coding Based on Butterworth and wavelet analysis theory to denoise and extract features in a BCI system which could be blocked by alpha, the dissertation design the control flow with feedback information to avoid false positive mistakes caused by EOG,EMG and other kinds of noise blocked by the alpha wave and obtain higher control accuracy of the mechanical leg. In addition, in the part of control though imaging humans’movements, the dissertation use many ways to extract the features such as the regression model(Auto Regressive, AR),wavelet transformation, Empirical mode decomposition(EMD) and combining AR and EMD theory. What’s more, the dissertation design the Fish linear classifier(FLD) and support vector machine(SVM) classifier to complete the design of the BCI system, and complete the preprocessing of EEG, feature extracting and pattern recognition though a large number of experiments. The study obtain a good experiments results, achieve the design of motor imagery BCI system and make a further attempt for the application that the BCI is used in the prosthetic field.