基于校验孪生神经网络的有线通信网络故障点定位方法研究  

Study on Fault Point Positioning Method of Wired Communication Network Based on Checking Twin Neural Network

作　　者：张强[1] ZHANG Qiang(Unit 61741 of PLA,Beijing 100094 China)

机构地区：[1]中国人民解放军61741部队,北京100094

出　　处：《长江信息通信》2024年第8期169-171,共3页Changjiang Information & Communications

摘　　要：由于有线通信网络中设备或者链路的承受能力有限,一旦发生故障将带来严重的负面影响,研究基于校验孪生神经网络的有线通信网络故障点定位方法。采集有线通信网络中的正常流量和故障流量数据,对采集到的流量数据进行清洗、变换等预处理操作,构建一个自校验的孪生神经网络模型,输入流量数据后故障点定位结果。实验结果表明,设计方法下有线通信网络故障点定位结果的准确度为98.65%,证实该方法的定位精度。Due to the limited bearing capacity of equipment or link in wired communication network,the failure will bring serious negative effects.The method of locating the fault point of wired communication network based on checking twin neural network is studied.Collect the normal traffic and fault flow data in the wired communication network,clean and transform the collected traffic data,build a self-verified twin ncural network model,and calculate the fault point location results after entering the traffic data.The experimental results show that the accuracy of the fault point positioning result of the wired communication network is 98.65%,which confirms the positioning accuracy of the design method.

关 键 词：校验孪生神经网络 有线通信网络 故障点定位 定位方法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]