基于压缩感知的无线通信网拓扑推断方法  

Topology Inference Method for Wireless Communication Networks Based on Compressed Sensing

作　　者：邵豪 王伦文 SHAO Hao;WANG Lunwen(Electronic Countermeasure 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,安徽合肥230037

出　　处：《探测与控制学报》2020年第2期92-98,共7页Journal of Detection & Control

基　　金：国防科技创新特区项目资助(17-H863-01-ZT-003-204-03);国家自然科学基金项目资助(61273302)。

摘　　要：针对无线通信非合作方难以使用传统拓扑发现方法获取网络拓扑的问题,提出基于压缩感知的无线通信网拓扑推断方法。该方法首先通过节点发出数据信号和确认信号的时间接续关系,获取时间窗口内网络节点状态;其次构造适用于无线通信网络的压缩感知模型框架,通过重构算法恢复节点链接向量;最后根据节点双向匹配原则法与筛选状态迭代法,筛选链路并提取相应时刻的节点状态,再次重构链接向量直至算法收敛。仿真实验表明,该算法能通过少量节点状态极化数据准确推断网络拓扑结构,具有较高时效性,且能够适应环境噪声干扰。To solve the problem that a non-cooperative party could not obtain the topology information of wireless communication networks by traditional topology discovery algorithm,a method to infer network topology by wireless signals was put proposed.Firstly,the status of nodes was obtained by the time-sequence relationships between data signal and reply signal.Secondly,a sparse vector recovery model was proposed,which was suitable for communication network by limited state sequences of nodes,and the link vectors were reconstructed.Finally,the correct states of links was extracted and the link vectors were reconstructed according to the bidirectional matching until the algorithm converged.The experiment results showed that the proposed algorithm could correctly infer the network topology with finite time series of the nodes.In addition,the algorithm had high timeliness and was less affected by the noise from environment.

关 键 词：拓扑推断 信号侦察 时间序列 压缩感知 向量重构 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]