基于自相关矩阵特征值的频谱盲感知方法

Blind Spectrum Sensing Algorithm Based on Eigenvalues of Autocorrelation Matrix

频谱感知是认知无线电的关键部分。为了提高频谱检测性能，提出利用接收信号平均功率与自相关矩阵最小特征值之比构建检测统计量来检测信号，该方法无需信号先验信息和噪声功率，是一种盲检测方法。以GPS信号为例，仿真表明该检测方法除了在较低信噪比时效果欠佳，整体检测性能优于能量检测方法，并降低了认知系统复杂度，适合于认知无线电系统。

Spectrum sensing is one of key parts for CR communication system, in order to improve the spectrum sensing performance, this paper presented a sensing algorithm employing the differences between av- erage power of received signals and the minimum eigenvalue of autocorrelation matrix; the algorithm is blind without knowing any priori information of the signal and noise power. Setting GPS signal as an example, the simtflation and performance analysis show that the performance of the proposed sensing algorithm is greatly outperforms the traditional energy detection algorithm and reduce the computational complexity of cognitive system as a whole except in the very low SNR environment, thus, the method is fit for CR system.