首页|期刊导航|铁路通信信号工程技术|基于AI的安全态势感知在铁路通信网中的应用研究

基于AI的安全态势感知在铁路通信网中的应用研究OA

Research on Application of AI-Based Security Situation Awareness in Railway Communication Networks

中文摘要英文摘要

提出一种基于深度学习的铁路通信网络安全态势感知系统,通过设计双向长短期记忆网络(Bidirectional Long Short-Term Memory,BiLSTM)与卷积神经网络(Convolutional Neural Network,CNN)的融合模型,实现多源异构数据的实时威胁检测.系统采用流式计算架构,可将处理延迟控制在120 ms以内.结合铁路实际应用情况,模拟仿真数据运行环境进行实验验证,结果显示整体检测准确率达98.7%、召回率为96.2%、F1分数为97.4%;其中,端口扫描攻击识别率达到99.2%,分布式拒绝服务攻击(Distributed Denial of Service,DDoS)检出率为97.8%,较传统检测方法提升了9.39%.此外,该系统可有效降低误报率5.3%、漏报率14.8%,将安全事件处置时效提升62%,为高速铁路通信系统提供了坚实的安全防护技术支撑.

This paper proposes a DL-based security situation awareness system for railway communication networks to achieve real-time threat detection based on multi-source heterogeneous data by designing a BiLSTM-CNN hybrid model.This system utilizes a stream processing architecture,which enables a processing latency within 120 ms.Based on the practical application scenarios of railway systems,experimental verification is conducted by simulating the operating environment with simulation data.The results show an overall detection accuracy of 98.7%,a recall rate of 96.2%,and an Fl score of 97.4%.Specifically,this approach offers a detection rate of 99.2%for port scanning attacks,and a detection rate of 97.8%for DDoS attacks,which represents an improvement of 9.39%compared to traditional detection methods.Furthermore,this system effectively reduces the false positive rate by 5.3%,and the false negative rate by 14.8%,while improving the handling efficiency of security incidents by 62%.Thus,it offers strong technical support for the security protection of communication systems for high-speed railways.

吴大勇;李慧;傅卫国

北京鼎兴达信息科技股份有限公司,北京 100070北京鼎兴达信息科技股份有限公司,北京 100070北京鼎兴达信息科技股份有限公司,北京 100070

交通工程

铁路通信网安全态势感知人工智能

railway communication networksecurity situation awarenessArtificial Intelligence

《铁路通信信号工程技术》 2026 (8)

77-82,6

10.3969/j.issn.1673-4440.2026.08.011

评论