当前位置:
X-MOL 学术
›
Radio Sci.
›
论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Spectrum Monitoring of Radio Digital Video Broadcasting Based on an Improved Generative Adversarial Network
Radio Science ( IF 1.6 ) Pub Date : 2021-07-23 , DOI: 10.1029/2021rs007270 X. Y. Wang 1 , J. J. Yang 1 , L. Zhang 1 , Q. N. Lu 1 , M. Huang 1
Radio Science ( IF 1.6 ) Pub Date : 2021-07-23 , DOI: 10.1029/2021rs007270 X. Y. Wang 1 , J. J. Yang 1 , L. Zhang 1 , Q. N. Lu 1 , M. Huang 1
Affiliation
Due to the inherent broadcast nature of wireless communication systems, instances of radio jamming are common, such as natural interference and man-made interference, resulting in increasing demands for radio monitoring. A spectrum monitoring method based on a generative adversarial network model that is one of the most promising approaches of learning any kind of data distribution using unsupervised learning was proposed in this paper for the detection of anomaly spectrum with impulse noise. To validate the performance of the proposed model, both the simulated data set and the measured data set of radio digital video broadcasting were used to train and test the model. Experiments on the two data sets reached a consistent conclusion: as long as the energy of the interference is greater than a certain threshold, the detection accuracy increases with the increase of the interference power and pulse width. Compared with the existing anomaly detection models, our model was faster and more stable.
中文翻译:
基于改进的生成对抗网络的无线电数字视频广播的频谱监测
由于无线通信系统固有的广播特性,无线电干扰的情况很常见,例如自然干扰和人为干扰,导致对无线电监测的需求不断增加。本文提出了一种基于生成对抗网络模型的频谱监测方法,该方法是使用无监督学习学习任何类型数据分布的最有前途的方法之一,用于检测具有脉冲噪声的异常频谱。为了验证所提出模型的性能,使用模拟数据集和无线电数字视频广播的实测数据集来训练和测试模型。在两个数据集上的实验得出了一致的结论:只要干扰的能量大于某个阈值,检测精度随着干扰功率和脉冲宽度的增加而增加。与现有的异常检测模型相比,我们的模型更快、更稳定。
更新日期:2021-08-07
中文翻译:
基于改进的生成对抗网络的无线电数字视频广播的频谱监测
由于无线通信系统固有的广播特性,无线电干扰的情况很常见,例如自然干扰和人为干扰,导致对无线电监测的需求不断增加。本文提出了一种基于生成对抗网络模型的频谱监测方法,该方法是使用无监督学习学习任何类型数据分布的最有前途的方法之一,用于检测具有脉冲噪声的异常频谱。为了验证所提出模型的性能,使用模拟数据集和无线电数字视频广播的实测数据集来训练和测试模型。在两个数据集上的实验得出了一致的结论:只要干扰的能量大于某个阈值,检测精度随着干扰功率和脉冲宽度的增加而增加。与现有的异常检测模型相比,我们的模型更快、更稳定。