一个面向AI识别任务的无线新型数据集

打开文本图片集

关键词：人工智能；数据集；无线感知；频谱监测；信号识别

doi：10.3969/J.ISSN.1672-7274.2025.11.006

中图分类号：TN92；TN98;TP18 文献标志码：B 文章编码：1672-7274（2025）11-0023-04

A Novel Dataset of Radio Signals for AI-Assisted Recognition Tasks

JUXinjue¹，PENG Jinbo²，XUHongliang1，³，ZHANG Yuxin²，CHEN Zhe²，GAO Yue² （1.Shanghai Radio Monitoring Station，Shanghai 2ooo31，China;2. Institute of SpaceInternet，FudanUniversity，Shanghai2oo433，China;3.CollegeofElectronic andInformation Technology，Tongji University，Shanghai 2018o4，China）

Abstract：AI-assistedapproaches forradio signal sensingandrecognitionserveas critical technologies inradio spectrum monitoring and integrated sensing and communication （ISAC）field.This paper systematicaly introduces a theoretical framework and construction methodologyforan AI-oriented radio signal dataset. Witha primary focus onelectromagnetic environments，thedataset objectively demonstrates signal characteristics across diverse complex scenarios，encompassing both singular and multiple sources under static and dynamic conditions.Experimental results demonstrate that AImodels trainedonthis dataset efectively accomplish signalrecognition tasks.This dataset significantly bridges the fundamental data gap inthe fieldof spectrum sensing，thereby providing essential data resources forAI-empowered spectrummanagement.

Keywords： artificial intelligence; dataset; radio sensing; spectrum monitoring; signal recognition

研究背景

无线感知作为连接物理与数字世界的关键技术，在频谱管理、低空经济及5G/6G等领域具有广泛应用。（剩余7230字）