无线电工程2024，Vol.54Issue(8)：1854-1861,8.DOI:10.3969/j.issn.1003-3106.2024.08.003
鲁棒多尺度神经网络的频谱感知方法研究
Research on Spectrum Sensing Method by Robust Multiscale Neural Network
摘要
Abstract
In Cognitive Radio(CR),Spectrum Sensing(SS)is a key technology that supports dynamic spectrum allocation and enhances spectrum utilization.However,traditional SS methods are susceptible to the impact of noise uncertainty,resulting in lower detection accuracy and a higher computational parameter load in low Signal to Noise Ratio(SNR)environments.To solve these problems,an SS method by Robust-Multiscale Neural Network(R-MsNN)based on Signal Processing(SP)features is proposed.This method combines the advantages of multiscale neural network and Gated Recurrent Unit(GRU)to effectively address the challenges faced in SS.The multiscale neural network gradually extracts abstract features from the bottom layer to the top layer to enhance the robustness of signal recognition.GRU selectively retains and forgets past temporal information,thereby better capturing long-term dependencies and temporal relationships.To validate the generalization capability of R-MsNN,the SS performance of R-MsNN is compared with that of different Deep Neural Network(DNN)architectures in two different noise models:one generated from a Generalized Gaussian Distribution(GGD)as noise model 1,and experimental data collected from unoccupied frequency modulation broadcasting channels as noise model 2.The experimental results demonstrate that,compared to the optimal model,R-MsNN trained with a combination of SP features achieved an average detection probability increase of 1.74％,2.55％,2.08％,and 1.59％for four different noise models 1,and a 1.72％increase for noise model 2.Additionally,compared to GRU,R-MsNN has half the number of parameters.This indicates that R-MsNN trained with a combination of SP features,exhibits robustness in a variety of complex noise en-vironments and can meet the dual requirements of high detection probability and low parameter number in SS tasks.关键词
频谱感知/信号处理/神经网络Key words
SS/SP/neural network分类
信息技术与安全科学引用本文复制引用
孟水仙,闫森,王树彬..鲁棒多尺度神经网络的频谱感知方法研究[J].无线电工程,2024,54(8):1854-1861,8.基金项目
国家自然科学基金(62361048)National Natural Science Foundation of China(62361048) （62361048）
