由于免许可证频段电磁环境复杂，因此LTE(Long Term Evolution)业务在该频段部署实现多业务共存的关键是对其进行监测和分类。为此，采用NS-3 网络仿真模拟器建立了LTE和Wi-Fi室内业务共存模型，通过改变网络和场景的各类参数构建了监测模型，获得了文件传输FT(file transfers)、音频流VF(voice flows)和恒定比特流CBRS(constant bit rate streams)三种模式下的数据集。结果表明，采用FT数据集，在K最近邻、支持向量机、决策树和随机森林算法的分类准确率有明显提高;同时，VF与CBRS两种数据集在上述算法上准确率达到了 80%左右。
Research on LTE Classification and Network Monitoring in Complex Environment
Due to the complex electromagnetic environment in the license-free band,monitoring and classification is the key issue for ensuring the deployment of LTE(Long Term Evolution)and its coexistence with Wi-Fi.To ad-dress this issue,a coexistence model for LTE and Wi-Fi indoor services is built using the NS-3 network simulator,and a monitoring model is constructed by varying various parameters of the network and scenarios to obtain the data-sets of FT(file transfers),VF(voice flows)and CBRS(constant bit rate streams)in three modes.Results show that the classification accuracy of K-nearest neighbors,support vector machines,decision trees and random forest algo-rithms has been significantly improved when using the FT dataset;meanwhile,an accuracy of about 80%can be a-chieved when using VF and CBRS datasets.
NETL
NSTL
万方数据
