Spectral estimation of phase noise based on complementary autocorrelation

In this paper, we propose a novel estimation technique¹ of the power spectral density (PSD) of phase noise based on complementary autocorrelation (CAC). Utilizing the fact that the CAC of phase noise is nonzero, this technique offers a distinct advantage that the PSD of phase noise is estimated while canceling other circular-symmetric noises. This property is most useful when estimating phase noise directly from digitally-modulated signals because they typically are accompanied by various kinds of circular noises that easily mask the phase noise of interest. A practical method to compute the phase noise PSD is given. In addition, an expression for a noise suppression factor of the method is given. Numerical simulations demonstrate the effectiveness of the proposed method, and verifies the noise suppression factor.