Detection of Signals in Nonstationary Random Noise: An Approach Using Evolutionary Spectrum and Stationarity Test

Abstract

In this paper, a novel method is proposed for detection of signals from the observation data corrupted by nonstationary random noise. Key approach is to estimate time-varying coefficients related as nonstationarity from the observation data and the observation data is transformed to stationarity process. As a method of signal detection, stationarity test is employed. Several numerical simulations show that our proposed method is effective to detect signals.