摘要
为加强通信信道的安全性,及时发现信道出现的故障和异常情况,该研究基于Petri网络模型设计出信道故障诊断系统,并利用时间标签和颜色集对Petri网络模型进行改进,使其更加全面表示出系统信道的动态行为;设计出动态探测器,与诊断模型共同进行故障诊断,采用FPGA与ARM结合的硬件设计方式,并加入了低运算放大器降低系统中的噪声;采用特征生成卷积神经网络构建信道故障预测模型,利用对抗生成网络学习故障样本的真实分布,并生成新的样本数据进行模型的训练;实验结果显示该研究系统的故障检测时间最短,故障预测准确率最高为99.7%。
In order to enhance the security of communication channel and timely detect faults and anomalies in a channel,a channel fault diagnosis system based on Petri network model is studied,and time labels and color sets are used to improve the Petri network model,so that the dynamic behavior of the system channel can be more comprehensively expressed.A dynamic detector is designed to diagnose faults together with the diagnostic model.The hardware design is the combination of FPGA and ARM,and a low operational amplifier is chosen to reduce the noise of the system.A feature generating convolutional neural network is used to construct a channel fault prediction model.An adversarial generation network is used to learn the real distribution of fault samples and generate new sample data for model training.The experimental results show that the system has the shortest fault detection time,and the highest fault prediction accuracy is 99.7%.
作者
朱铝芬
徐媛媛
ZHU Lüfen;XU Yuanyuan(Intelligent Manufacturing Institute,Nanjing Vocational College of Information Technology,Nanjing 210046,China)
出处
《计算机测量与控制》
2023年第6期6-11,共6页
Computer Measurement &Control
基金
国家自然科学基金项目(21KJD470004)。
关键词
Petri网络模型
信道故障诊断
动态探测器
低运算放大器
特征生成卷积
故障预测
petri network model
channel fault diagnosis
dynamic detector
low operational amplifier
feature generation convolution
failure prediction