Improving SNR estimation for linearly modulated signals over SIMO channels via oversampling

It is important to accurately estimate the signal-to-noise ratio (SNR) without the a priori knowledge of modulation type and transmitted symbols in certain particular areas of communications. In this paper, we introduce the idea of oversampling and propose two novel families of second-order and fourth-order statistics based estimators. It is shown by Monte Carlo simulations that both the two families of SNR estimators clearly outperform the typical estimator in [7] and are robust to phase and synchronization uncertainty. The fourth-order statistic based estimators also show robustness to large frequency offset. The superiority and feasibility of the proposed estimators makes them suitable to be exploited in non-cooperative communications.