高斯色噪声背景下的单通道信源数目估计算法

目前信源数目估计算法大都是基于多通道接收模型且对高斯色噪声抑制能力较差，而实际应用中单通道接收模型及色噪声环境非常普遍，因此研究色噪声背景下的单通道信源数目估计算法意义重大。针对现有算法的缺陷提出了一种基于构建信号时间快拍和四阶累积量矩阵的单通道信源数目估计算法。首先通过构建信号时间快拍实现单通道接收信号的升维得到矢量化空间，然后以此组信号空间构造出四阶累积量矩阵，并从理论上验证了该四阶累积量矩阵能有效抑制高斯白噪声及高斯色噪声的影响，最后对该矩阵进行奇异值分解并通过信息论准则估计出信源个数。仿真实验和实际信号实验都表明本文算法能较好地解决单通道信源数目估计问题，且能有效抑制高斯色噪声。 

Single Channel Source Number Estimation Algorithm Under Colored Gaussian Noise

At present,most source number estimation algorithms are based on multi-channels receive model and have weak performance to restrain colored Gaussian noise.However,single channel model and color noise are common in real life application.So it is of great importance to research single channel source number estimation algorithm under colored Gaussian noise.In view of limitation of algorithm existed,a novel single channel source number estimation algorithm based on construction of time snapshots and fourth-order cumulant matrix is presented.Firstly,time snapshots theory is applied to increase signal dimensions of single channel received signal and vectorization spaces are obtained.Then these sets of spaces are adopted to construct fourth-order cumulant matrix.It is validated theoretically that the proposed fourth-order cumulant matrix can effectively restrain white and colored Gaussian noise.Finally,singular value decomposition(SVD) is applied on this matrix and source number can be estimated by information theoretic criteria.Simulation and real-life experiment results both indicate that the proposed algorithm can be applied to single channel source number estimation and has well performance to restrain colored Gaussian noise.