多径信道下基于EM算法的盲LDPC编码器识别研究  

EM-based blind LDPC identification in multipath channels

作　　者：刘钰 王方刚 张静文 艾渤[1] 钟章队[1] LIU Yu;WANG Fanggang;ZHANG Jingwen;AI Bo;ZHONG Zhangdui(State Key Lab of Rail Traffic Control and Safety,Beijing Jiaotong University,Beijing 100044,China)

机构地区：[1]北京交通大学轨道交通控制与安全国家重点实验室,北京100044

出　　处：《通信学报》2018年第9期43-48,共6页Journal on Communications

基　　金：中央高校基本科研业务费专项基金资助项目(No.2018JBM078);国家自然科学基金资助项目(No.61571034;No.61725101);北京自然科学基金资助项目(No.4182051;No.L172020);泛网无线通信教育部重点实验室基金资助项目(No.KFKT-2018102);北京科委重大专项基金资助项目(No.Z181100003218010)~~

摘　　要：随着认知无线电的出现,盲编码器识别引起了广泛的关注。现有相关工作主要集中在加性高斯白噪声(AWGN)信道,而多径场景下的盲识别尚未得到充分研究。考虑未知多径衰落信道下的盲低密度奇偶校验码(LDPC)识别,提出了基于期望最大化(EM)算法的似然分类器,以获得未知参数的最大似然估计值,并采用平均对数似然比(LLR)估计器对未知编码器进行识别。数值结果表明,该算法在多径信道中具有良好的识别性能,特别是在低信噪比区域。As the advent of cognitive radios,blind encoder identification has attracted increasingly attentions since it plays an important role.The existing works mainly focus on additive white Gaussian noise(AWGN)channel,while the blind identification in multipath scenarios has not been sufficiently investigated.Considering the blind low density parity-check(LDPC)codes identification in the presence of unknown multipath fading channel,a likelihood-based classifier was proposed using the expectation maximization(EM)algorithm to obtain the maximum likelihood estimates of the unknown parameters.Then,an average log-likelihood ratio(LLR)estimator was adopted to classify the unknown encoder.Numerical results show that the proposed algorithm provides promising identification performance in multipath channels,especially in the low signal-to-noise ratio region.

关 键 词：认知无线电 多径信道 盲编码器识别 LDPC EM算法 

分 类 号：TN929[电子电信—通信与信息系统]