Empirical Bayesian Method in the Problem of a Harmonic Signal Detecting in the Gaussian Noise

The problem of detecting a harmonic signal with unknown parameters against the background of uncorrelated Gaussian noise of unknown power is considered. The expediency of using the empirical Bayesian method on the basis of estimates of their parameters, including the use of reconstructed values of the latent signal, is substantiated. Adaptive signal detection algorithms under different conditions of a priori uncertainty are synthesized for two adequate models of harmonic signal - trigonometric and autoregressive. Comparative studies of the efficiency of synthesized detectors were performed by the Monte Carlo method and proved the superiority of detectors synthesized by the autoregressive model