In-service SNR estimation without symbol timing recovery for QPSK data transmission systems

Cramér-Rao Bounds for SNR Estimation of Oversampled Linearly Modulated Signals

“…A few authors have considered SNR estimation with oversampling. Seeking robustness to symbol timing uncertainty, moments-type NDA SNR estimators based on the oversampled MF output were proposed in [22]. The DA ML SNR estimator and corresponding Cramér-Rao Bound (CRB) based on the oversampled signal at the MF input, with arbitrary oversampling factor, were given in [9].…”

Estimation of carrier and channel parameters in time‐selective fading channels

“…This holds also true for channel parameters, most prominently expressed by estimation of the SNR [2]. It was the rising interest in adaptive schemes during the last decade, which caused that powerful SNR algorithms have been developed for all relevant categories: (a) data-aided or non-data-aided [11,12], depending on whether data symbols are known to the receiver or not; (b) coherent or non-coherent with respect to carrier frequency/ phase [13], depending on the possibility of successful carrier synchronisation; (c) based on baudrate or oversampling [14], depending on whether the symbol timing could be reliably established in advance or not. Governed by powerful error correction schemes such as turbo and LDPC schemes, code-aware parameter estimation was extensively explored over the last decade as well [15][16][17].…”