无线电工程2025，Vol.55Issue(7)：1431-1439,9.DOI:10.3969/j.issn.1003-3106.2025.07.008
无线电频谱占用度测量及信号检测方法
Research on Radio Spectrum Occupancy Measurement and Signal Detection Methods
摘要
Abstract
Radio frequency occupancy is a key indicator that characterizes the real occupation status of radio frequency spectrum.In response to the issues in current national standards for measuring radio frequency occupancy and the research on the relationship between frequency occupancy and wireless channel state,a radio signal transmission and spectrum reception system is established.The relationship between radio frequency occupancy measurement methods and channel noise models is investigated.Experimental results show that if the channel is a Gaussian noise,the national standard method can be used for frequency occupancy measurement.If the channel is a non-Gaussian noise,the 5-sigma principle is more suitable for statistical frequency occupancy measurement.Additionally,machine learning methods are applied to investigate the wireless channel occupancy state.The results indicate that when the Signal to Noise Ratio(SNR)is greater than 15 dB,the classification accuracy of five types of radio signals based on a signal spectrum classification model can exceed 99％.Improving the classification accuracy of radio signals under low SNR and non-Gaussian noise channels remains a challenge.关键词
频谱占用度/信号检测/5西格玛原则/机器学习方法Key words
spectrum occupancy/signal detection/5-sigma principle/machine learning methods分类
天文与地球科学引用本文复制引用
张婕,杨晶晶,黄铭,董城愿..无线电频谱占用度测量及信号检测方法[J].无线电工程,2025,55(7):1431-1439,9.基金项目
国家自然科学基金(62261059,62361055)National Natural Science Foundation of China(62261059,62361055) （62261059,62361055）
