Nonlinear signal reconstruction based on the decomposition into chaotic components

The paper proposes a signal reconstruction technique based on the decomposition into chaotic components. The considered approach can be usefully associated with the filtering, forecasting and control algorithms when only a small number of data samples is available. The developed decomposition algorithm involves sequential component extraction and recursive computation of the cost function. Some related questions are also discussed: choice of the class of chaotic maps, computational complexity of parameter estimation.