The adaptive optimal differentiation by standard deviation criterion

The method is suggested to solve the differentiation problem. It allows building the estimates for the derivatives of Gaussian stationary signals that are close to the optimal ones by the standard deviation criterion when the spectral density of a useful signal and a noise are known to within level. The problem is solved by the specifically designed nonlinear dynamic systems. A nearly optimal solution of the differentiation problem is found when the equivalent transfer function parameters of nonlinear dynamic differentiator can be made close to the parameters of the optimal Wiener filter for any fixed level of rational spectral density of a useful signal and a noise.