基于深度学习的矿井无线网络流量预测研究

Research on mine wireless network traffic prediction based on deep learning

为准确预测矿井无线网络流量变化情况,确保井下无线网络安全稳定运行,保障矿井安全生产。在分析无线网络流量特征的基础上提出一种时空卷积全连接网络(CL-FCCNet),它是基于残差网络(ResNets)和循环神经网络(RNN)的无线网络流量预测模型。预测模型能够针对复杂的无线网络工作环境进行流量预测,及流量数据中存在的时空特征,帮助实现流量监控异常自动报警。试验结果表明:模型的预测效果较现有预测方法具有一定的提升。

In order to accurately predict the change of mine wireless network flow,ensure the safe and stable operation of underground wireless network and ensure the safety of mine production.Based on the analysis of the traffic characteristics of wireless network,this paper proposes a kind of CL-FCCNet,which is a traffic prediction model based on residual network(ResNets)and circular neural network(RNN).The prediction model can forecast traffic for complex wireless network working environment,model the temporal and spatial characteristics of traffic data,and help to realize the abnormal automatic alarm of traffic monitoring.The experimental results show that the prediction effect of the model is improved compared with the existing prediction methods.