Comparing performances of adaptive filter methods for model independent noise cleaning of chaotic signals

In this thesis, Gaussian noise added chaotic signalsare filtered using different adaptive filtering algorithms, that utilize the known noise variance only, without any knowledge of system models of chaotic signals, and performances of these filtering algorithms are compared. The filtering process is done by using Universal Finite Impulse Response Minimum Mean Square Error filter (Universal FIR MMSE filter), the Wiener filter and LMS-like filter,filter designed using a Polynomial Fitting method and the filter designed using Wavelet Coefficient Shrinkagemethod. Using Matlab programming language, performances of the implemented filters are compared. Finally, the correlation dimensions of noise-free chaotic signals and those of filteroutput signals are calculated using Matlab and compared.