Performance analysis of SNR estimators in digital receivers

Abstract

The performances of the maximum likelihood (ML) signal-to-noise ratio (SNR) estimators developed in a companion paper [NC Beaulieu et al, 2005] are examined. Approximations to the probability density functions of the ML SNR estimates in a sampled signal processing receiver as well as a continuous time signal processing receiver are derived. Two measures of signal-to-noise ratio are studied in detail. The results are general and apply to both static additive white Gaussian noise channels and slowly fading channels.