基于本征非线性的非合作无线发射源个体识别机制研究

Considering the contradiction between spectrum demand and limited available spectrum resource in the complex radio environment, the illegal use of spectrum and malicious attacks seriously threaten the spectrum security. The existing radio monitoring technologies are not applicable to effectively support the detection and tracking of interference sources. On this base, by extracting the differences from wireless channels and physical layer information of radio devices, wireless emitter identification, identifies where the received signal originate from. It attracts lots of attention and becomes a key metric of measuring the capabilities of network information security. However, there is a certain gap between our country's research in this field and that in the advanced and developed countries. Hence it is imperative to conduct relevant research. On this base, this project will research on the contradiction between the speed and accuracy of the radiometric identification. Through the cognition of a specific communication emitter, nonlinear modeling of the transmitter system will be constructed. Through the cognition of the received signal, the multi-dimensional characterization of the received signal can be conducted. Through the cognition of the spectrum environment, the extracted feature engineering is processed and the strong classifier is designed. Moreover, the deeply mining engine is constructed to optimize the speed and accuracy of the radiometric identification, so as to realize the rapid and accurate radiometric identification under complex electromagnetic environment and support the increasingly diverse radio monitoring tasks.

在频谱供需矛盾日益突出的今天，非法用频现象与恶意攻击行为严重影响了频谱使用秩序，危害频谱安全。现有的无线电监管技术难以有效支撑干扰源排查任务。对此，基于信号射频与物理层特征的“无线发射源个体识别”，立足于发射源本征非线性、利用传播信道的差异性特征，并通过接收信号直接判断其来源于哪个特定发射机，提供干扰源目标持续追踪能力，受到广泛关注并已成为衡量频谱安全保障能力的关键指标。然而，我国在此方面研究与先进发达国家相比有一定差距，开展相关研究势在必行。本项目将研究无线发射源个体识别速度与准确性之间的矛盾问题，通过对无线发射源个体认知，实现对发射机本征非线性建模；通过对接收信号认知，实现对信号多维度描述与表征；通过频谱环境认知，对信号特征进行针对性优化并设计强分类、强泛化能力的分类器。最终结合构建的深度挖掘引擎，在复杂电磁环境下实现对非合作无线发射源的快速、准确辨识，支撑日趋多元的无线电监管任务。

本项目面向无线通信网络安全认证中关键技术-“通信发射源个体识别”，开展了三年持续研究工作，聚焦于无线发射源个体识别速度与准确性之间的矛盾这一科学问题，重点研究了“无线发射机非线性表征问题”、“限定条件下的动态排序问题以及限制条件集最小化问题”、与“多维特征复杂性与有限时间分析准确性问题”三方面的理论问题。具体地，通过对通信发射源个体认知，实现对发射机系统非线性建模；通过对无线信号认知，实现对接收信号多维度描述与表征；通过频谱环境认知，对信号特征进行针对性优化并设计有效分类器。开发了基于软件无线电平台的半实物仿真验证系统，结合构建的深度挖掘引擎，在复杂电磁环境下实现对通信发射源快速、准确辨识。依托本项目发表SCI检索论文10篇，IEEE会议论文6篇，其他会议论文8篇，申请国家发明专利9项，获授权发明专利6项，申请PCT专利3项，申请计算机软件著作权5项，项目成果获得2021年中国通信学会科技一等奖提名。项目成果应用于北京2022年冬奥会测试赛无线电安全保障工作和新疆维吾尔自治区边境电磁环境常态化监测活动中，有效扩大了无线电监测范围，极大地节约了无线电监测的时间与人力成本。依托本项目协助培养博士研究生1名，硕士研究生7名。