基于贝叶斯优化XGBoost的协作频谱感知算法  

COOPERATIVE SPECTRUM SENSING ALGORITHMBASED ON BAYESIAN OPTIMIZED XGBOOST

作　　者：胡延飞 郭滨[1] 孙佳楠 Hu Yanfei;Guo Bin;Sun Jianan(College of Electronic Information Engineering,Changchun University of Science and Technology,Changchun 130022,Jilin,China)

机构地区：[1]长春理工大学电子信息工程学院,吉林长春130022

出　　处：《计算机应用与软件》2024年第7期121-127,共7页Computer Applications and Software

基　　金：吉林省科技发展计划项目(20200404216YY)。

摘　　要：为了提高无线信道环境的频谱感知性能,提出一种基于贝叶斯优化XGBoost的协作频谱感知算法。在一个主用户(PU)和三个次用户(SU)的协作频谱感知场景下,提取信号的归一化能量特征,采用贝叶斯优化算法同时优化XGBoost模型的多个超参数,最后利用优化XGBoost算法实现待检测信号的分类。仿真结果表明,与传统频谱感知算法和KNN、GNB、SVM、MLP等机器学习算法相比,该算法在Rayl和AWGN信道环境检测准确率分别为88.4%和90.25%,可以有效提高不同信道环境下的协作频谱感知性能。In order to improve the spectrum sensing performance of the wireless channel environment,a cooperative spectrum sensing algorithm based on Bayesian optimization XGBoost is proposed.In a cooperative spectrum sensing scenario of a primary user(PU)and three secondary users(SU),the normalized energy characteristics of the signal were extracted.The Bayesian optimization algorithm was used to optimize multiple hyperparameters of the XGBoost model at the same time,and the optimized XGBoost algorithm was used to realize the classification of the signal to be detected.The simulation results show that compared with traditional spectrum sensing algorithms and machine learning algorithms such as KNN,GNB,SVM,MLP,the detection accuracy of this algorithm under Rayl and AWGN channel are 88.4%and 90.25%,respectively,which can effectively improve the cooperative spectrum sensing performance in different channel environments.

关 键 词：认知无线电 频谱感知 贝叶斯优化 XGBoost 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]