一种平坦衰落信道下的盲信噪比估计算法

针对基于子空间分解信噪比估计算法中信号子空间维数估计复杂度高、小样本条件下估计偏差大的问题,提出了一种改进的盲信噪比估计算法.该算法首先构造接收信号的自相关矩阵,然后从矩阵奇异值序列的尾部开始,间隔两项依次进行差分得到梯度序列,再以梯度序列相邻两项均值大干特定阈值为条件确定信号子空间的维数,最后求得信噪比.仿真结果表明:信噪比范围为-5～+15 dB时,平坦衰落信道下常用调制信号的信噪比估计标准差小于0.1 dB,与MDL,AIC方法相比,该算法计算量小,且能适应更低的信噪比和更短的数据长度.

A Blind SNR Estimation Algorithm over Flat Fading Channel

In order to solve problems of the complexity of estimating dimension of signal subspace and the large deviation of SNR estimation under little data length in the algorithm of SNR estimation based on subspace decomposition. An efficient algorithm is proposed to estimate SNR over flat fading channel. Firstly, a autocorrelation matrix of received signal must be constructed, then we can obtain the gradient array from the end of the singular value array by getting deviation every three values in order, and the dimension of signal subspace will be acquired on condition that the average of two consecutive elements of the gradient array is greater than specific threshold, finally we can obtain SNR estimation. Computer simulations show that the standard deviations for different modulation signals over flat fading channel are less than 0.1dB when the actual SNR ranges from -5dB to 15dB, not only that, this algorithm performs better in lower SNR and shorter data samples environment compared with MDL and AIC method.