认知无线电中一种非高斯背景下的协同频谱感知方案

A cooperative spectrum sensing scheme under the background of non-Gaussian in cognitive radio

在认知无线电系统中,由于需要对一定频谱范围内的认知用户进行频谱检测,而传统经典的检测方法无法满足复杂多变的检测环境,基于此,文章研究了在实际情况中更为普遍的非高斯噪声情况下认知用户的频谱检测问题.由于传统的广义似然比检测计算量较大而且实际工程中较难实现,文章采用广义似然比的一种近似形式即Rao检测.通过采用矩估计方法对系统模型的混合系数等进行参数估计,并在拟合非高斯噪声的背景下应用于认知用户的频谱检测,推导出非高斯噪声中Rao检测的检测统计量和检测性能公式,分析比较了非高斯噪声中Rao检测与高斯Rao检测的性能.仿真结果表明,在虚警概率一定的情况下,基于Rao检测的频谱检测方法能有效地提高认知无线电网络中非高斯噪声背景下认知用户的检测性能.

In cognitive radio system,due to the need for cognitive users within a certain range of the spectrum for spectral detection but the classic detection methods cannot meet the complex testing environment.For this reason,this paper studies the spectrum sensing problem which more cognitive users in non Gauss noise generally under in actual situation.Because the traditional generalized likelihood ratio test which the large amount of calculation and difficult to implement in practical engineering,this paper uses generalized likelihood ratio test ＇s approximate forms,that is Rao detection.By using the method of moments estimate mixing coefficient system model parameter estimation and fitting spectrum detection applied to cognitive users under non- Gaussian noise background,deducing the non-Gaussian noise test statistic Rao detection and detection performance formula,analysis and comparison of the non-Gaussian Rao noise detection and Gaussian Rao detection performance.The simulation showed that,under certain circumstances the probability of false alarm,spectrum sensing method based on Rao detection can effectively improve the detection performance of cognitive users under No-Gauss noise background in cognitive radio networks.