APPLICATION OF THE SPECTRAL METHOD TO STOCHASTIC FILTER ANALYSIS

The problem of standing out a signal from an additive mixture of a harmonic signal and white Gaussian noise is considered. The analysis is based on the phenomenon of stochastic resonance (SR), which consists in amplifying a periodic signal under the influence of noise of a certain power. SR is a universal physical phenomenon that is typical of some nonlinear systems, and is came out not only in technical, but also in biological and social systems. When calculating the spectral characteristics of the output signal, Volterra series were used. The problem is solved using the transfer functions of Volterra without the initial definition of kernels. Volterra transfer functions are obtained by the harmonic input signal method. The influence of the input signal parameters, in particular the amplitude and frequency of the harmonic signal and the noise power, on the spectral power density of the output signal is studied. Optimal parameters values are determined. Criteria are formulated for using a stochastic filter to standing out a harmonic signal on the background white Gaussian noise.