基于改进机器学习的无人机网络入侵自动感知系统设计

Design of an Automatic Intrusion Sensing System for Unmanned Aerial Vehicle Networks Based on Improved Machine Learning

提升无人机网络的安全通信能力,是保证无人机应用的基础,因此,设计基于改进机器学习的无人机网络入侵自动感知系统。系统数据模块通过NetFlow网络流量采集器,采集网络的内部安全流量,管控模块管理并调用这些数据传送至入侵检测模块中,该模块中的入侵感知网关通过部署的时空卷积网络模型,检测无人机网络入侵行为,并完成入侵行为分类,全面感知无人机网络通信状态;感知结果则通过可视化模块进行展示。测试结果显示:该系统能够精准检测网络入侵行为并及时发送入侵预警,网络中主机的内核安全程度均在0.955以上,保证网络的安全通信。

Improving the secure communication capability of drone networks is the foundation for ensuring drone applications.Therefore,a drone network intrusion automatic perception system based on improved machine learning is designed.The system data module collects internal security traffic of the network through the NetFlow network traffic collector,and the control module manages and transfers this data to the intrusion detection module.The intrusion detection gateway in this module detects drone network intrusion behavior through the deployed spatiotemporal convolutional network model,completes intrusion behavior classification,and comprehensively perceives the communication status of the drone network;The perceptual results are displayed through visualization modules.The test results show that the system can accurately detect network intrusion behavior and send intrusion warnings in a timely manner.The kernel security level of hosts in the network is above 0.955,ensuring safe communication of the network.