A Nonlinear Method For Signal Processing Based On Coupled Map Lattice Model

Electroencephalography, for short EEG, is an external representation of human brain's behaviors, and also an important assistant tool for clinical diagnosis. While event-related potentials (ERP) are called "a window for observing brain's advanced functions" for their consanguineous relationship with human's psychological activities. But because of the limitation of classical EEG analysis, most EEG analysis lacks catholicity. Attempts to use it as a clinical diagnostic tool has not yet successful and been accepted by psychiatrist. The purpose of this thesis is to study and discuss the EEG's characteristics using nonlinear dynamics. Since brain wave has proved to be typical nonlinear and non-stationary, at the same time, ERP signals can be regarded as low-dimensional chaos, we think it is more appropriate to analyze ERP dynamical characteristics with nonlinear and non-stationary methods.In this dissertation, we firstly reviewed the developmental process and study actuality about ERP and nonlinear dynamics introduced the coupled map lattice model frequently used in nonlinear dynamics domain. Then we built several model and compared their abilities about simulating actual system's activities by computer emulation. Aiming at the problems in models, we raise some measures to correct them, and the results proved that the new models have better ability to simulate actual system's activities. EEG signals are accepted as representative non-stationary time series. But because of the limitation of their study methods, lots of researchers were based on the hypothesis that EEG is (quasi) stationary. At present time, the non-stationary study of the typical non-stationary signals is primary. We calculated the time-variant maximal Lyapunov exponent of ERP using time-variant coupled map lattice and calculated the global correlation dimension using mended GP algorithm. We obtained some quantized parameters which can reflect the global characters of system and discover information that traditional methods can not do. And the discovery and use about this information will make more achievement in clinical diagnose and Physiological Psychology.