الفهرس 
قرار المناقشة والحكمOnly 14 pages are availabe for public view
 |  | from | 145 |  |  |
| | | from | 145 | | |
Abstract
Communication between people is very important and more complex than any other form of communication, and plays an important role in any relationship. Similarly, as artificial devices become more complicated and play an important role in everyday life, communicating effectively with people and devices becomes increasingly important. Human i.nteracts with the world using his five senses and limb movements. These senses and limbs are controlled by the brain, so the study of the brain signal is very important in order to simulate the action of the brain signal. .~ People who are suffering from amputated arm need alternative methods for communication and control. The Brain Computer Interface (BCI) is the unique so lution till now. In this work, I present new approaches in analysing (EEG) signals, generated by the brain, which can be used in controlling the movements of an artificial arm. Researchers have analysed EEG signals in the time domain by extracting features such as mean, standard deviation average, ect. Some researchers have used the analysis of the discrete and continuous wavelet analysis. However, non-stationary characteristics of EEG signals has a valuable impact in improving the classification of signal and the mean of the absolute value of signal. While the non- linear features are extracted from a density matrix which is generated from the phase space of the signal and from the recurrence quantification analysis (RQA). The density matrix method features were extracted by using the gray level eo-occurrence matrix (GLCM). However, the recurrence quantification analysis features were extracted by using the nonlinear recurrence plots toolbox. Four classification approaches have been applied; the linear discriminant analysis (LDA), support vector machine (SVM), Bayes and KNN classifiers. The Graz 2003 datasets has been used in this work. The maximal achieved classification rate is 90% by the combination of the amplitude frequency analysis and recurrence quantification analysis. The results confirmed the robustness of the new tecaaique and demonstrate its value as a classification approach in the field of brain computer interface BCl. This thesis is a great step towards designing of a real time recognition system that can read the real brain signal and extract the optimal features and classify it to right or left classes that can be used as a control signal to control a bionic arm.