基于多维特征融合的鲁棒频谱感知技术  

Multi-dimensional Fusion Feature Based Robust Spectrum Sensing Scheme

作　　者：母诗源 王晓庆[1] MU Shi-yuan;WANG Xiao-qing(The 54^th Research Insttitute of CETC,Shijiazhuang Hebei 050081)

机构地区：[1]中国电子科技集团公司第五十四研究所,河北石家庄050081

出　　处：《数字技术与应用》2020年第12期23-25,共3页Digital Technology & Application

摘　　要：在存在不确定性噪声的环境中,传统的频谱感知方法存在一定的局限性。文中提出基于多维特征融合的鲁棒频谱感知技术,运用样本协方差矩阵的特征值信息进行频谱感知,实现较好的抗干扰能力。基于多维特征值作为检验统计量,通过高斯近似,根据奈曼-皮尔逊准则导出判决阈值的闭式表达式。即使没有授权信号和衰落信道的先验知识,也能得到很好的检测性能。基于大量的仿真结果表明,该技术不受噪声影响,与其他现有方法相比,具有更好的检测性能。In uncertain noise environment,the performance of traditional spectrum sensing methods have many limitations.In this paper,by using the eigenvalues of the sample’s covariance matrix,a robust spectrum sensing scheme based on multi-dimensional fusion feature is proposed,which achieves better anti-interference ability.Specifically,test statistics are constructed by multi-dimensional eigenvalues,and the closed form decision threshold is derived via Gaussian approximation based on the Neyman-Pearson criterion.Even without the prior knowledge of primary user’s signal and fading channel,the proposed method performs well.Additionally,extensive simulations are provided,which show that the proposed method can counter the adverse effects of noise uncertainty,and outperforms other existing methods.

关 键 词：认知无线电 频谱感知 多维特征融合 

分 类 号：TN925[电子电信—通信与信息系统]