基于改进DBSCAN的电台通联关系发现方法

Research on network connection relationship based on improved DBSCAN

摘要:

针对目前无线战术网络拓扑主要通过协议解析的方法感知，对于非合作复杂场景下实现难度大的问题，提出一种基于频谱信号的物理特征和统计规律，采用改进的DBSCAN算法的通联关系挖掘方法；通过分析网络中频谱信号的通信特点，确定了选择载波频率、带宽、跳频周期和平均信号功率等特征作为信号标签进行通信行为的挖掘，利用改进的DBSCAN算法，自适应选择合适聚类参数对数据进行处理；通过对聚类结果的分析实现目标通联关系的判断；实验结果证明，该方法在没有解析通信双方协议的前提下，能够成功辨识通信节点的个体行为和通信关系等重要信息，为挖掘和发现节点的通联关系提供了新的研究思路，具有较高的工程应用价值。

Abstract:

In view of the fact that the current wireless tactical network topology is mainly perceived through protocol analysis, which is difficult to achieve in non-cooperative complex scenarios, a communication relationship mining method based on the physical characteristics and statistical laws of spectrum signals and an improved DBSCAN algorithm is proposed; by analyzing the communication characteristics of spectrum signals in the network, it is determined to select characteristics such as carrier frequency, bandwidth, frequency hopping period and average signal power as signal labels for communication behavior mining, and the improved DBSCAN algorithm is used to adaptively select appropriate clustering parameters to process the data; the target communication relationship is judged by analyzing the clustering results; experimental results show that this method can effectively discover important information such as individual behavior and communication relationship of communication nodes without deciphering the content carried by the spectrum signals, which provides a new research idea for mining and discovering the communication relationship of nodes and has high engineering application value.