SVD降噪在载波同步技术中的研究

为了提升突发通信中载波频偏估计的性能,降低有效估计的信噪比门限,提出了一种基于奇异值分解去噪的频偏估计方法。首先,该方法将含噪信号根据相应映射转换成矩阵,并对该矩阵进行奇异值分解;其次,将较大的奇异值判定为信号特征予以保留,否则判定为噪声特征置零;接着,根据降噪后的奇异值重构矩阵,恢复成模拟信号;最后,将预处理后的信号进行M&M频偏估计。结果表明,相比于不去噪的频偏估计算法,该方法能够提升估计精度,降低信噪比门限,具有一定的实用性。

Research of SVD in Carrier Synchronization Technology

In order to improve the performance of carrier frequency offset estimation algorithms in burst-mode communication, and reduce the SNR threshold in effective estimation, a new frequency offset estimation method based on singular value decomposition is proposed. Firstly, this method transforms the primal signal into matrix via a relevant reflection, and then implements singular value decomposition of the matrix. Secondly, the several fairly large singular values are determined as signal characteristic for reservation or otherwise as noise characteristic turning into zero. Thirdly, based on the new singular values the analog signals are reconfigured .Finally, M&M frequency offset estimation is done on the preprocess signals. Experiment results show that, as compared with traditional algorithm, the proposed method could raise the estimation precision, reduce the SNR threshold, and thus is of certain practical value.