频谱感知是认知无线电网络中重要的、不可或缺的一部分.近年来,随着随机矩阵理论的发展,把随机矩阵理论应用到认知无线电中,提出了很多能改善感知性能的算法,使频谱感知的性能得到显著提升.因此,基于矩阵论研究了一种基于矩阵特征值上下界的感知算法(BED),通过判定所有特征值是否在判决界区间内来判断频谱是否空闲,然后充分利用矩阵的所有特征值,根据估计的噪声方差和M-P律设定判决门限,分别在不同的采样数、协作用户数、信噪比下进行仿真比较.理论推导和仿真结果均表明,新型感知算法性能明显优于未利用全部特征值的其他算法,检测概率显著提升,感知性能得到优化.%Spectrum sensing is the most important and integral part of cognitive ratio networks. In recent years, with the development of random matrix theory and the application of random matrix theory to cognitive radio. Lots of algorithms are proposed to improve the perception performance and make the spectrum-sensing performance receive a significant boost. According to matrix theory, a perceptron algorithm based on the upper and lower bounds of matrix eigenvalues (BED) is explored. Whether the spectrum is free is decided by determining whether all eigenvalues are within the bounds of the decision, and then by fully utilizing all the eigenvalues of the matrix, and the decision threshold is set in accordance with the estimated noise variance and the M-P law. Simulations and comparisons are done respectively at different sampling number, cooperative-user number and SNR. Experiment indicated that this novel algorithm is clearly superior to the other traditional algorithm in detection probability and perception performance.
展开▼
