| Olfaction is one of the most essential sense system, keeping daily living and life of people and animals. At present, we know little about olfactory physiological knowledge, which is disproportion with the rich and colorful olfaction. The olfactory study can not only help us understand its mechanism better, but also has the important value for studying olfactory bionics and comprehending signal mechanism of brain and other sense system.Firstly, in this paper, physiological structure of olfactory system is introduced and its model's study status quo is summarized. The compositive structures of K set model brought forward by Freeman are mainly analyzed and simulated, through which the olfactory EEG is gotten. Furthermore, the development and application of K set model are summarized.On the basis of obtaining the simulated olfactory EEG, the signal characteristics during the stimulation are analyzed by a few parameters in the time domain, frequency domain and non-linear aspect. The chaotic parameters such as correlation dimension and largest Lyapunov exponent are mainly discussed. The results show it's feasible and proper to apply K set model to study olfactory system and to get signal characteristics of olfactory EEG with these signal analytical methods.Besides, the simulated signals and mixed EEG from the MIT-BIH database are respectively separated by two kinds of representive independent component analysis (ICA) methods--extended Infomax ICA and fast fix-point ICA based on the kurtosis, through which these two methods are compared. Then, the mixed simulated olfactory EEG is separated by the latter one. The results show that it's feasible to extract olfactory EEG with ICA.At last, The mixed EEG with several kinds of noises and the mixed simulated olfactory EEG with noises are respectively denoised and separated by the method combining the wavelet denoising and ICA. The results show that the combination of these two methods is effective for analyzing mixed olfactory EEG with noises. |
PDF Full Text Request