The Awgn Channel Snr Estimation Algorithm

This paper introduces typical four SNR estimation algorithms. Then three SNR estimation algorithms, which are high-order cumulants estimator, singular value decomposition of signal self-correlation matrix estimator and data fitting estimator, are studied over AWGN channel. At last, we assess and compare the estimators' performance under an identical simulated model using identical performance metric-the mean square error (MSE), and derive the Cramer-rao bound (CRB).Monte Carlo simulation results show that the estimation errors of high-order cumulants estimator and singular value decomposition of signal self-correlation matrix estimator can be controlled within 0.6dB when the actual SNR ranges from -10 to 20dB; and the estimation error of data fitting estimator is less (?)an 0.3dB when actual SNR ranges from -5 to 12dB. The normalized MSE of singular value decomposition of signal self-correlation matrix estimator is less than 0.03, and the normalized MSE of other two estimators is close to the CRB.