一种采用刀切法的单通道信源数估计算法

针对现有信源数估计算法不能直接用于单通道接收模型且抑噪能力较差的问题，提出了一种采用刀切法的单通道信源数估计算法。该算法首先通过间隔抽样实现了单通道接收信号多维数的转换，得到矢量化空间；然后采用刀切法将此组空间重构多个协方差矩阵，经酉变换后结果取平均；最后通过循环迭代得到最优信源数。理论分析和仿真结果表明，该算法在白、色噪声环境下能有效抑制噪声，且在低信噪比及采样点较少时能更准确估计信源数，相较于传统的估计算法，显著提高了检测性能。

A single channel source estimation algorithm based on Jackknife method

In order to solve the problem that the existing source number estimation algorithms can not be directly employed in reception of signals for single channel and have poor noise suppression capability,a single channel source estimation algorithm based on Jackknife method is proposed.Firstly,the multi-dimensional conversion of single-channel received signal is realized by interval sampling method,and the vectorized space is obtained.Then,the space is reconstructed into multiple covariance matrixes by Jackknife method and the results are averaged after unitary transformation.Finally,the optimal number of sources is obtained by loop iteration.Theoretical analysis and simulation results show the algorithm can effectively suppress noise both in white noise and color noise environment,and can estimate the number of sources more accurately under conditions of low signal-to-noise ratio(SNR) and few sampling points.Compared with the traditional estimation algorithm,the detection performance is improved remarkably.