一种数字通信信号盲信噪比估计方法

摘要

针对数字通信信号的信噪比盲估计问题,提出了一种基于子空间理论的盲信噪比估计方法.该方法首先根据信号自相关序列构建特定维数的信号自相关矩阵,并根据实际工程应用需求,利用坐标旋转数字计算(Coordinate rotation digital computer,CORDIC)算法实现Jacobi旋转来完成自相关矩阵的特征值分解,避免了实际实现时对硬件乘法器的调用.并以6种常用的数字通信信号为例,在加性高斯白噪声(AWGN)信道条件下,实际信噪比在-10～30 dB范围内时对其信噪比估计性能进行仿真分析.仿真表明,当实际信噪比为-5～22 dB时,信噪比估计标准偏差小于0.5 dB,且提出的信噪比估计器具有渐近无偏特性,证明了该方法是一种进行盲信噪比估计的有效方法.%A subspace theory based approach to blind signal-to-noise ratio (SNR) estimation for digital communication signals is proposed. Firstly, the method constructs the autocorrelation matrix with special dimensions according to the autocorrelation sequence of the signals, and then considering its real applications, coordinate rotation digital computer (CORDIC) algorithm is used to implement Jacobi rotation to autocorrelation matrix eigenvalues decomposition, which can avoid the use of multipliers in its hardware implementation. Moreover, six kinds of digital modulated signals are applied to simulations for blind SNR estimations. Performance of the estimator is simulated and analyzed in additional white noise (AWGN) channel while the simulation SNR is set as-10-30 dB. Finally, simulations show that the estimation of standard deviation is less than 0. 5 dB when the actual SNR is-5-22 dB, and the proposed SNR estimator has unbiased asymptotic properties, which proves that the method is effective to blind SNR estimation.