The Two-Step Non-Data-Aided SNR Estimation in the Low SNR Region of OFDM Signals

摘要

In this paper, a novel two-step non data-aided (NDA) signal to noise ratio (SNR) estimator is proposed to improve its accuracy in the low SNR region of orthogonal frequency division multiplexing (OFDM) signals. The two-step estimator consists of a coarse estimation followed by a refinement step. In the first step, three linear independent coarse estimations of the signal and noise power are obtained by exploring the cyclic-prefix (CP)-induced redundancy. In the second step, these three coarse estimations are refined by resorting to the best linear unbiased estimator (BLUE). Compared with the schemes which only rely on two out of these three coarse estimations, a more accurate SNR estimation can be obtained in the low SNR region. Simulation results show that the proposed noise and signal power estimator achieve approximately 6dB and 2dB SNR gain respectively in the low SNR region, therefore, the proposed SNR estimator provides approximately 4dB SNR gain.