基于深度学习的电子通信设备信号异常识别研究设计  

Research design of signal anomaly identification in electronic communication equipment based on deep learning

作　　者：康晋[1] KANG Jin(Yangling Vocational&Technical College,Yangling ShaanXi 712100,China)

机构地区：[1]杨凌职业技术学院,陕西杨凌712100

出　　处：《自动化与仪器仪表》2025年第3期228-232,共5页Automation & Instrumentation

基　　金：陕西省教育厅2023年度科学研究计划项目《信息化背景下大学生职业素质能力构建与研究》(23JK0263)。

摘　　要：针对当前信息化背景下电子通信设备信号异常识别方法存在识别精度低,识别效率不高的问题,提出一种基于深度学习的电子通信设备信号异常识别方法。首先,以小波包方法进行能量谱的特征提取,并采用旁路滤波方法对小波包进行改进,通过其解决小波包方法分析振动信号的频谱混叠问题,以提升信号特征提取精度;然后将提取特征输入至基于卷积神经网络(Convolutional Neural Network,CNN)的信号异常识别模型中,通过CNN网络实现信号异常准确识别。结果表明,本模型进行信号异常识别的平均精确率和平均召回率分别为94.68%和96.34%,均高于传统的BP(Back Propagation)模型和支持向量机(Support Vector Machine,SVM)模型,且本模型的信号异常识别时长仅为11.35 s,对比于另外两种模型分别降低了30.95 s和15.51 s。由此说明,本模型能够提升电子通信设备信号异常识别的精确率和效率,满足通信行业中通信设备信号异常识别需求,具备优越性。In view of the problems of low identification accuracy and low identification efficiency of electronic communication equipment under the current information background,a signal abnormal identification method of electronic communication equipment based on deep learning is proposed.Firstly,the energy spectrum is extracted by the wavelet packet method,and the wavelet packet is used to improve the problem by solving the wavelet packet method to improve the signal feature extraction accuracy;then the extracted features are input to the signal abnormality identification model based on convolutional neural network(Convolutional Neural Network,CNN),and realize accurate signal identification through CNN network.The results show that the average accuracy and average recall rate of signal anomaly identification in this model are 94.68% and 96.34%,respectively,which are higher than the traditional BP(Back Propagation) model and support vector machine(Support Vector Machine,SVM) model,and the signal anomaly identification time of this model is only 11.35 s,which reduces 30.95 s and 15.51 s compared with the other two models.This shows that this model can improve the accuracy and efficiency of abnormal signal identification of electronic communication equipment,meet the requirements of abnormal signal identification of communication equipment in the communication industry,and has advantages.

关 键 词：深度学习 小波包 特征提取 CNN网络 异常识别 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]