Research on SVD SNR Estimation Algorithm

Signal-to-noise ratio (SNR) is an important indicator of communication quality and an indispensable priori information in the signal processing process. In this paper, the SNR estimation algorithms based on singular value decomposition (SVD) are analyzed from two different perspectives. They are the matrix construction method and SNR calculation characteristics. At the same time, the advantages and disadvantages of these different algorithms are discussed. The simulation models are established to explain the influence of two matrix construction methods on SNR estimation and the accuracy and robustness of four SNR estimation methods. The simulation results show that the performance of Hankel matrix is more stable and the ratio determination method is more accurate and with lower complexity.