【摘要】 为了提高频谱感知性能,克服经典算法的缺点,提出了一种新的基于Wishart随机矩阵理论的协作频谱感知算法。根据多个认知用户接收信号样本协方差矩阵特征值的对数分布特性,利用样本协方差矩阵最大特征值与几何平均特征值的比值,得到简单的判决阈值闭式表达式,实现频谱感知判决。该算法不需要知道主用户的任何先验信息,不受噪声不确定性的影响。仿真结果表明,所提算法在协作用户数少、信噪比低、采样点数极少的情况下,仍能获得较高的感知性能。该算法受虚警概率和极端值的影响较小,比同类算法有更好的检测性能。更多还原

【Abstract】 In order to improve the spectrum sensing performance and overcome the shortcomings of the classical algorithm, a new cooperative spectrum sensing algorithm based on Wishart random matrix theory was proposed. According to the logarithmic distribution characteristics of the sampled covariance matrix eigenvalues and using the ratio of maximum eigenvalue and geometric mean eigenvalue, a simple closed-form threshold expression could be obtained, and the spectrum sensing decision could be performed depend on the threshold. The simulation results show that the proposed algorithm can get better sensing performance even under the conditions of a few number of cooperative users, low signal to noise ratio and a few samples. It is less affected by false-alarm probability and the extreme values, and has better detection performance than similar algorithms.更多还原