Entropy-Based Non-Data-Aided SNR Estimation

Abstract

A non-data-aided SNR estimation of a communication channel is addressed in this paper. The main contribution relies on the relationship between an entropy measure of a given constellation and the link quality. The main advantage is that the second order Rényi entropy is invariant to the rotation of the constellation and it is not envelope-based as the moment-based methods, enabling to asymptotically approximate the CRLB of coherent estimators at high SNR for an increasing data size. Comparative results are shown with other non-data-aided SNR estimation techniques.