基于核空间优化支持向量机的合作频谱感知算法  

Cooperative Spectrum Sensing Algorithm Based on Kernel Space Optimization Support Vector Machine

作　　者：南亚飞 张云蕾 朱芮 Nan Yafei;Zhang Yunlei;Zhu Rui(School of Microelectronics,Tianjin University,Tianjin 300072,China)

机构地区：[1]天津大学微电子学院,天津300072

出　　处：《南开大学学报（自然科学版）》2021年第3期8-14,共7页Acta Scientiarum Naturalium Universitatis Nankaiensis

摘　　要：在认知无线电网络中,由于噪声不确定性引起的聚类重叠,导致能量检测的性能明显降低.针对噪声不确定对频谱感知的影响,提出了一种基于核空间优化支持向量机的合作频谱感知算法.该算法融合了支持向量机和核空间优化相关理论,将感知用户收集的数据统计量组合成向量,使用Fisher准则对该向量集进行相关运算,得出使各类数据在高维空间中分离度最高的核函数参数.之后使用支持向量机算法对训练向量进行训练,得到最优的频谱感知分类器.仿真结果表明,基于核空间优化支持向量机的合作频谱感知算法在噪声不确定情况下优于传统的合作频谱感知算法.In cognitive radio networks,clustering overlap due to noise uncertainty results in significantly reduced energy detection performance.Aiming at the influence of noise uncertainty on spectrum sensing,a cooperative spectrum sensing algorithm based on kernel space optimization support vector machine is proposed.The algorithm combines the theory of support vector machine and kernel space optimization,combines the statistics collected by the perceptual users into vectors,and uses Fisher’s criterion to perform correlation operations on the vector sets,so that the resolution of various types of data in high-dimensional space is obtained.The highest kernel function parameter.Then the training vector is trained using the support vector machine algorithm to obtain the optimal spectrum perceptual classifier.The simulation results show that the cooperative spectrum sensing algorithm based on kernel space optimization support vector machine is superior to the traditional cooperative spectrum sensing algorithm in the case of noise uncertainty.

关 键 词：合作频谱感知 支持向量机 核空间优化 

分 类 号：TN923[电子电信—通信与信息系统]