非线性时间序列在卫星通信网络数据预测中的研究

Research on Satellite Communication Network Data Forecast System Based on Non-linear Time Series

卫星通信网络数据预测在卫星通信系统中起着重要的作用,是系统建模的主要研究内容之一。由于卫星通信网络数据的非平稳性和不可预知因素的影响,决定了应采用非线性时间序列建模方法来分析、预测。在分析通信网络数据的基础上,建立卫星通信网络数据的AR IMA模型,在确定预测模型的阶和进行参数估计后,给出不同预测步数条件下的通信网络数据流量的预测,并进行了仿真对比实验。仿真结果表明,该模型在预测步数较小的情况下,预测误差在4%左右,具有良好的预测精度,为卫星通信网络数据流量的预测、异常检测和网络负载预测的应用奠定了坚实的基础。

Satellite communication network data forecast system which is one of the main studies in system modeling plays an important role in Satellite communication system. Because the Satellite communication network data is not stationary and there are unpredictable factors, nonlinear time series modeling method should be adopted to analyze and forecast it. Based on the analysis of the data communication network, the ARIMA model is established. When the order of the prediction model is determined and parameter estimation is done, forecasts of communication network data flow under the conditions of different forecast step are given, and comparison simulation experiments are carried out. Simulation results show that, the model＇s forecast error is around 4% in predicting the smaller step, so it has good prediction accuracy and provide a solid foundation for satellite communications network data flow forecast, anomaly detection and network load forecast.