«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年03期

页码:

180-186

栏目:

人工智能

出版日期:

2023-03-10

文章信息/Info

Title:

Single User Spectrum Sensing Algorithm Based on Kernel Space Optimization SVM

文章编号:

1673-629X(2023)03-0180-07

作者:

余 飞¹ ; 岳文静¹ ; 陈 志²

1. 南京邮电大学 通信与信息工程学院,江苏 南京 210023;

2. 南京邮电大学 计算机学院,江苏 南京 210023

Author(s):

YU Fei1 ; YUE Wen-jing1 ; CHEN Zhi2

1. School of Telecommunications & Information Engineering,Nanjing University of Posts & Telecommunications,Nanjing 210023,China;

2. School of Computer,Nanjing University of Posts & Telecommunications,Nanjing 210023,China

关键词:

核空间优化; 支持向量机; 小波降噪; 被囊群算法; 单用户频谱感知

Keywords:

nuclear space optimization;  support vector machine;  wavelet noise reduction;  tunicate swarm algorithm;  single userspectrum sensing

分类号:

TP301. 6;TN391. 41

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 03. 027

摘要:

在认知无线电领域中,由于噪声随机动态变化引起信号聚类重叠,导致能量检测性能较差,为了解决能量检测效率低以及噪声变化对频谱检测性能的影响,提出了一种基于核空间优化 SVM 的单用户频谱感知算法。 该算法将支持向量机和核空间优化相关理论相结合,通过对信号频谱占用以及空闲两种状态构建出认知信号,对信号进行小波降噪处理后,构建出特征向量进行训练和学习,从而得到判断频谱状态的分类模型,并利用自适应 t 分布变异策略以及萤火虫扰动算法对被囊群算法寻优过程进行改进和加速,优化训练搜索得到最佳核函数参数 滓 和惩罚系数 C 。 仿真实验结果表明,提出的基于核空间优化支持向量机的单用户频谱感知算法和传统的能量检测以及协作频谱感知算法相比较,具有较高的检测准确性和鲁棒性

Abstract:

In the field of cognitive radio,the random dynamic change of noise causes signal clustering overlap,resulting in poor energy detection performance. In order to solve the low efficiency of energy detection and the impact of noise changes on the performance ofspectrum detection,a single user spectrum sensing algorithm based on nuclear space optimization SVM is proposed,which combines thetheory of support vector machine and nuclear space optimization,construct a cognitive signal by occupying the signal spectrum and idletwo states, and performs wavelet noise reduction treatment on the signal. The eigenvectors are trained and learned to obtain aclassification model for judging the spectral state,and the adaptive t-distribution variation strategy and the firefly perturbation algorithmare used to improve and accelerate the optimization process of the tunicate swarm algorithm,and the optimal kernel function parameter and punishment coefficient C are obtained by optimizing the training search. Simulation results show that the proposed single userspectrum sensing algorithm based on the nuclear space optimization SVM has relatively high detection accuracy and robustness comparedwith the traditional energy detection and collaborative spectrum sensing algorithm.

常用功能

更新日期/Last Update: 2023-03-10