Use of chaotic modeling for transmission of EEG data

A method based on chaotic modeling for compression and transmission of EEG data is presented. It is assumed that the data to be transmitted is chaotic and is generated by an E-dimensional nonlinear system. The flow in the E-dimensional space is obtained from the one dimensional time series using the process of time delay embedding. Then a simple, linear model is fitted for this flow in the E-dimensional space and this model is used as a predictor in the general DPCM scheme for transmission. Since this model was able to give very good one-step prediction, we were able to achieve a reduction of about 50 percent of the dynamic range of the signal to be transmitted (and hence a reduction in bit rate also). A further advantage is that this is achieved with the additional feature of reduced complexity when compared to the conventionally used AR model.