基于Friedman检验的非参数协作频谱感知方法

摘要

Covariance matrix based spectrum sensing encounters performance degradation when there the antenna correlation is low. To overcome this drawback, a nonparametric cooperative spectrum sensing algorithm based on Friedman test is proposed. Distributed sensors possess the effect of space diversity, so that the signal power among the sensors at the same time may not be completely equal. Based on this feature, the spectrum sensing is realized by comparing signal powers among the sensors. For the nonparametric approach is adopted, the proposed algorithm is robust to noise uncertainty and is suitable for noise of any statistical distribution. The theoretical expression of decision threshold is also derived, which shows that the decision threshold has no relationship with the sample number. As a result, the threshold does not need to be reset when the sample number changes. Simulation results demonstrate the effectiveness of the algorithm.%协方差矩阵频谱感知方法在天线相关性低时感知性能较差，该文针对这一问题提出一种基于Friedman检验的非参数协作频谱感知方法。分布式放置的感知节点具有空间分集的特性，因此在同一时刻感知节点上的信号功率不完全相同。利用这一特点，提出通过比较各感知节点的信号功率水平来实现频谱感知。由于采用了非参数化表示，该方法对噪声不确定性稳定，且适用于任意统计分布的噪声。另外，推导了所提方法判决门限的理论表达式，结果显示判决门限与采样点数无关，因此在采样点数变化的情况下无需重新设置判决门限。仿真结果验证了上述理论分析的有效性。