SNR estimation via oversampling for SIMO communcations with linear modulation

It is required to accurately estimate signal-to-noise ratio (SNR) without the a priori knowledge of modulation types and transmitted symbols in certain areas of communications. In this paper, we propose second-order and fourth-order statistics based estimators by oversampling the received signal. The expressions of the estimators under different scenarios are derived. From computer simulation, both families of SNR estimators outperform the typical M4 estimator and are robust to phase and synchronization uncertainty. The superiority and feasibility of the proposed estimators make them suitable to non-cooperative communications networks.