| 引用本文： |
-
张行,庄理淇,季飞,魏冬,黄伟庆,付婧雯.基于上行控制信道的移动通信空口流量分析技术[J].信息安全学报,已采用 [点击复制]
- zhanghang,zhuangliqi,jifei,weidong,huangweiqing,fujingwen.Uplink Control Channel based Traffic Analysis Technique for Mobile Communication Airports[J].Journal of Cyber Security,Accept [点击复制]
|
|
| 摘要: |
| 为了深入评估移动通信空口开放性导致的信息泄漏问题，本文提出基于上行控制信道的移动通信空口流量分析技术，该技术利用协议中混合自动重传请求（Hybrid Automatic Repeat Request，HARQ）的反馈数据在物理层的时频占用特征实现精细化的流量分析。相较于传统的基于空口链路层的流量分析技术，该技术不需要先验信息，也不需要解析数据，并且在空口信令加密的情况下，仍然可以完成流量分析。针对多终端混合的现实场景，该技术基于移动通信上行控制信道时频资源占用实现用户数目的识别以及用户数据的分离，并在分离数据的基础上，提取时频占用的统计特征，利用机器学习算法实现空口流量的精细化分析。首先，该技术利用控制信道一个调度周期单次调度用户的特点，实现监测时间内的用户数量识别，并基于该调度周期内的功率数据生成用户数据分离算法的基准；在此基础上，利用不同用户由于信道衰落、通信距离等因素导致的功率差异，实现用户数据的分离；最后，通过提取分离数据的时频占用统计特征实现不同终端通信业务的识别。针对本文提出的流量分析技术，我们基于开源的srsRAN平台搭建了模拟4G和5G环境，并在实验环境下，对用户数目识别、用户数据分离以及用户业务识别算法进行了验证。实验结果表明，用户数目识别的准确率达到了95%，不同业务的识别准确率超过了96%。此外，我们采用不同品牌的手机验证了本文算法的普适性。 |
| 关键词: 移动通信空口安全 上行控制信道 流量分析 混合数据分离 机器学习 |
| DOI： |
| 投稿时间：2024-04-12修订日期：2024-07-30 |
| 基金项目:国家重点研发计划 |
|
| Uplink Control Channel based Traffic Analysis Technique for Mobile Communication Airports |
|
zhanghang, zhuangliqi, jifei, weidong, huangweiqing, fujingwen
|
| (Institute of Information Engineering,CAS) |
| Abstract: |
| This paper proposes a technique for analyzing mobile communication air port traffic based on the uplink con-trol channel. The technique utilizes the time-frequency occupancy characteristics of the feedback data of Hy-brid Automatic Repeat Request (HARQ) in the protocol in the physical layer to realize fine-grained traffic analysis. The objective of this paper is to evaluate the information leakage problem caused by the openness of mobile communication air ports. This technology uses the time-frequency occupancy characteristics of the HARQ feedback data in the physical layer to achieve precise traffic analysis. Unlike traditional traffic analysis techniques based on the air interface link layer, this technique does not require prior information or parsed data. It can still complete traffic analysis even when air interface signaling encryption is present. This technol-ogy enables the identification of user numbers and separation of user data based on the time and frequency re-source occupation of the uplink control channel of mobile communication, allowing for refined analysis of air interface traffic using machine learning algorithms. It is designed to address the realistic scenario of mul-ti-terminal mixing. The statistical characteristics of the time and frequency occupation are extracted from the separated data. The language used is clear, concise, and objective, with a formal register and precise word choice. The text adheres to conventional structure and formatting features, including consistent citation and footnote style. The grammar, spelling, and punctuation are correct. No changes in content have been made. The technique utilizes the characteristics of a single scheduling user in one scheduling cycle of the control channel to identify the number of users during the monitoring time. It generates the benchmark of the user data sepa-ration algorithm based on the power data in this scheduling cycle. The language is clear, objective, and val-ue-neutral, and avoids biased, emotional, figurative, or ornamental language. The sentence structure is simple and the technical terms are consistent. The text is free from grammatical errors, spelling mistakes, and punc-tuation errors. No changes in content were made as per the instructions. Based on these factors, the system leverages power differences resulting from channel fading, communication distance, and other variables to separate user data. This is achieved by extracting statistical characteristics of time and frequency occupation of the separated data. Finally, machine learning algorithms are used to perform a detailed analysis of the air interface traffic. The paper proposes a traffic analysis technique that identifies different terminal communica-tion services by extracting time-frequency occupancy statistical characteristics from separated data. The al-gorithms for subscriber number identification, subscriber data separation, and subscriber service identifica-tion were verified in an experimental environment using simulated 4G and 5G environments based on the open source srsRAN platform. The experimental results indicate that the accuracy of user number recognition is 95%, and the accuracy of different service recognition exceeds 96%. Furthermore, we validate the universality of the algorithms presented in this paper by testing them on different brands of cell phones. |
| Key words: Security for mobile communication air port Uplink control channel Traffic analysis Hybrid data separation Ma-chine learning |
