Estimation of Spectrum Parameters of Random Processes on the Basis of Observations in Noise

The case is considered in which the observed process is the sum of two independent stationary random processes, namely, the signal that is of interest and distortion noise, under the assumption that both processes are Gaussian. The paper considers the problem of testing the hypothesis that noise is present in the available observation data; a method is proposed for obtaining asymptotically effective estimates of the unknown spectral-density parameters of the signal and the unknown intensity of “white noise.”