基于频谱数据端到端的电台通联关系研究

End-to-end Radio Communication Relationship Recognition based on Spectrum Monitoring Data

对无线网络中通联行为进行识别和分析,对反恐维稳等国家安全具有重要意义。通联关系是挖掘无线网络中行为关系和网络中的隐藏信息的基础。通过识别通联关系,可以推测电台在通信网络中的层级位置,进而找到网络中的关键节点。端到端是指不需要人工设计专家特征,使用深度学习的方法直接通过频谱监测数据识别通联关系。经过仿真试验发现,残差神经的网络结构可以有效识别通联关系,在测试集上识别率达到99.02%。

The identification and analysis of communications in wireless networks is of great significance to national security such as anti-terrorism and stability maintenance.The communication relationship is the basis for mining the behavior relationship in the wireless network and the hidden information in the network.Through the identification of the communication relationship,the hierarchical position of the radio station in the communication network can be inferred,and then the key nodes in the network can be found.Endto-end means that there is no need to manually design expert features,and deep learning methods are used to directly identify communication relationships through spectrum monitoring data.Simulation experiments indicate that the residual neural network structure can effectively identify the communication relationship,and the recognition rate on the test set can reach 99.02%.