摘要：To evaluate the wireless network condition on moving medial vehicles and make reasonable judgment among multiple network channels, an automatic network judgment method is presented for unstable communication environment based on adaptive network speed analysis and prediction method. Historical network observation data are collected; time series analysis method and ARMA model are employed to build prediction model using historical data statistics with periodical update mechanism. Predicted network speed value is then compared with real speed value and a criterion to identify the condition of network is summarized from comparison result analysis. Experiments are conducted in both normal network and interfered network situations. Empirical studies results show that the adaptive prediction method could be used for wireless network speed prediction on moving vehicles. Under different network environments, the fit value of each group of data varies and this value could be used for network problem identification.
关键词：wireless network; ARMA model; time series analysis; data prediction
