基于拉普拉斯先验算法的多维度频谱感知  

Multidimensional Spectrum Sensing Based on Laplace Prior Algorithm

作　　者：俞启明 金虎 刘轶凡 郑文庆 YU Qiming;JIN Hu;LIU Yifan;ZHENG Wenqing(Electronic Warfare Institute,National University of Defense Technology,Hefei 230037,China)

机构地区：[1]国防科技大学电子对抗学院,合肥230037

出　　处：《电讯技术》2018年第4期370-375,共6页Telecommunication Engineering

摘　　要：全双工认知无线电LAT(Listen-and-Talk)模型能够在次用户传输数据的同时进行频谱感知,这要求感知用户能在次用户自干扰的影响下检测当前是否存在主用户。利用全局功率谱模型的方法实现了对当前用户的使用频段、发射功率和位置坐标的多维度感知,以此辨别该用户的身份。在求解数学模型时,选取了拉普拉斯近似函数用作未知参数的先验概率密度函数,与现有方法相比,提升了算法收敛速度,并获得了准确性更高的结果。The full duplex cognitive radio Listen-and-Talk(LAT)protocol enables secondary users per-form spectrum sensing and data transmission simultaneously.It requires sensing users to make reliable deci-sion on whether the primary user is transmitting under the influence of secondary user.Global power spec-trum model is applied to perform spectrum sensing multi-dimensionally,which can get the location of the current user and its signal's frequency and power.These information can identify the current user.When solving the math model,Laplace prior function is used as the prior probability density function of unknown parameter.Laplace prior algorithm outperforms current algorithms in speed and accuracy.

关 键 词：全双工 认知无线电网络 协作式频谱感知 全局功率谱模型 

分 类 号：TN924[电子电信—通信与信息系统]