MED和MME频谱感知算法的改进

Improvement on MED and MME Spectrum Sensing Algorithms

基于随机矩阵理论的最新研究成果,建立了由滤波缺陷和过采样引起的相关噪声下的频谱感知模型,分析采样协方差矩阵最大特征值的极限分布,推导了该模型下最大最小特征值(maximum-minimum eigenvalue,MME)和最大特征值检测(maximum eigenvalue detection,MED)算法的判决阈值公式.仿真结果表明:相关噪声下采用判决阈值的MME和MED算法的虚警概率与预设虚警概率吻合,而采用原MME和MED算法中的判决阈值的虚警概率则偏离预设虚警概率.

Based on the latest research result of random matrix theory, a spectrum sensing model of correlation noise caused by filtering defects and oversampling is established. The limit distribution of the maximum eigenvalue of the covariance matrix is analyzed, and the closed form threshold formula of the maximum-minimum eigenvalue （MME） and the maximum eigenvalue detection （MED） is deduced under the correlated noise model. The simulation results show that the false alarm probability of MME and MED deduced by the threshold under the correlated noise matches the predetermined false alarm probability. However, the false alarm probability of the original MME and MED deviates from the predefined false alarm probability.