TítuloCramér-Rao Bounds for SNR Estimation of Oversampled Linearly Modulated Signals
Tipo de publicaciónJournal Article
Year of Publication2015
AutoresLópez-Valcarce, R, Villares, J, Riba, J, Gappmair, W, Mosquera, C
JournalIEEE Trans. Signal Processing
Start Page1675
Date Published04/2015
Palabras claveadaptive signal processing, compass, Cramér-Rao bound (CRB), dynacs, oversampling, Signal-to-noise ratio (SNR)
Resumen Most Signal-to-noise ratio (SNR) estimators use thereceiver matched filter output sampled at the symbol rate, anapproach which does not preserve all information in the analogwaveform due to aliasing. Thus, it is relevant to ask whetheravoiding aliasing could improve SNR estimation. To this end, wecompute the corresponding data-aided (DA) and non-data-aided(NDA) Cramér-Rao bounds (CRBs). We adopt a novel dual filterframework, which is shown to be information-preserving undersuitable conditions and considerably simplifies the analysis. It isshown that the CRB can be substantially reduced by exploitingany available excess bandwidth, depending on the modulationscheme, the SNR range, and the estimator (DA or NDA) type.