基于似然比检验的超宽带信道检测与定位算法  

UWB Channel Detection and Location Algorithm Based on Likelihood Ratio Test

作　　者：孙希延[1,2] 刘健[1,2] 纪元法[1,2] 廖桂生[1,2] 范广伟 

机构地区：[1]桂林电子科技大学信息与通信学院,桂林541004 [2]广西精密导航技术与应用重点实验室,桂林541004 [3]卫星导航系统与装备技术国家重点实验室,石家庄050000

出　　处：《电子与信息学报》2017年第3期590-597,共8页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(61362005;61561016);广西自然科学基金(2013GXNSFA019004;2014GXNSFAA118352)~~

摘　　要：超宽带(Ultra Wide Band,UWB)室内定位系统的定位性能主要受信号非视距(None Line Of Sight,NLOS)传播影响。为此该文提出一种基于信道统计量(Channel Statistics Information,CSI)的信道NLOS状态检测法。该方法首先在IEEE802.15.4a信道模型下对均方根时延扩展和平均超量延迟的概率分布函数进行建模,作为信道标准分布。再以信道瞬时分布与标准分布间的KL散度为检验统计量做似然比检验(Likelihood Ratio Test,LRT)来鉴别信道状态。同时提出一种基于LRT的定位算法LRT-Chan算法。该算法能有效利用受NLOS污染的测距数据提高定位精度。仿真结果表明:LRT信道状态检测法在全部UWB信道中都能获得较高检测准确率;在定位锚点(Anchor Node,AN)分布不理想的NLOS环境中LRT-Chan算法也能取得较高定位精度。The performance of UWB indoor positioning system is mainly affected by NLOS errors. In this paper, a channel state detection method based on channel statistics is proposed. The probability distribution function of Root Mean Square Delay Spread（RMS） and Mean Excess Delay（MED） under IEEE802.15.4a standard is modeled as standard distribution. Channel state is identified by Likelihood Ratio Test（LRT） based on KL divergence between channel instantaneous distribution and standard distribution. A localization algorithm named LRT-Chan based on LRT is proposed to improve positioning accuracy by effectively utilizing data contaminated by NLOS. Simulation results show that, LRT detection can obtain high accuracy in all UWB channels; when Anchor Nodes（ANs） with NLOS errors are not in ideal distribution, LRT-Chan algorithm can gain higher positioning accuracy.

关 键 词：似然比检验 NLOS信道鉴别 LRT-Chan算法 KL散度 

分 类 号：TN92[电子电信—通信与信息系统]