提出了一类用于非线性时间序列建模的混合自回归滑动平均模型(MARMA).该模型是由K个平稳或非平稳的ARMA分量经过混合得到的.讨论了MARMA模型的平稳性条件和自相关函数.给出了MARMA模型参数估计的期望极大化(expectation maximization)算...提出了一类用于非线性时间序列建模的混合自回归滑动平均模型(MARMA).该模型是由K个平稳或非平稳的ARMA分量经过混合得到的.讨论了MARMA模型的平稳性条件和自相关函数.给出了MARMA模型参数估计的期望极大化(expectation maximization)算法.运用贝叶斯信息准则(Bayes information criterion)来选择该模型.MARMA模型分布形式富于变化的特征使得它能够对具有多峰分布以及条件异方差的序列进行建模.通过两个实例验证了该模型,并和其他模型进行比较,结果表明MARMA模型能够更好地描述这些数据的特征.展开更多
The non-stationary behavior, caused by the train rmverrent, is the rmin factor for the variation of high speed railway channel. To measure the tirce-variant effect, the parameter of stationarity interval, in which the...The non-stationary behavior, caused by the train rmverrent, is the rmin factor for the variation of high speed railway channel. To measure the tirce-variant effect, the parameter of stationarity interval, in which the channel keeps constant or has no great change, is adopted based on Zhengzfiou- Xi＇an （Zhengxi） passenger dedicated line measurement with different train speeds. The stationarity interval is calculated through the definition of Local Region of Stationarity （LRS） under three train ve- locities. Furthermore, the time non-stationary characteristic of high speed pared with five standard railway channel is corn- Multiple-Input MultipleOutput （MIMO） channel models, i.e. Spatial Channel Model （SCM）, extended version of SCM （SCME）, Wireless World Initiative New Radio Phase II （WINNERII）, International Mobile Teleconmnications-Advanced （IMT-Advanced） and WiMAX models which contain the high speed moving scenario. The stationarity interval of real channel is 9 ms in 80% of the cases, which is shorter than those of the standard models. Hence the real channel of high speed railway changes more rapidly. The stationarity intervals of standard models are different due to different modeling methods and scenario def- initions. And the compared results are instructive for wireless system design in high speed railway.展开更多
文摘提出了一类用于非线性时间序列建模的混合自回归滑动平均模型(MARMA).该模型是由K个平稳或非平稳的ARMA分量经过混合得到的.讨论了MARMA模型的平稳性条件和自相关函数.给出了MARMA模型参数估计的期望极大化(expectation maximization)算法.运用贝叶斯信息准则(Bayes information criterion)来选择该模型.MARMA模型分布形式富于变化的特征使得它能够对具有多峰分布以及条件异方差的序列进行建模.通过两个实例验证了该模型,并和其他模型进行比较,结果表明MARMA模型能够更好地描述这些数据的特征.
基金Acknowledgements This work was supported partially by the Beijing Natural Science Foundation under Crant No. 4112048 the Program for New Century Excellent Talents in University under Gant No. NCET-09-0206+4 种基金 the National Natural Science Foundation of China under Crant No. 60830001 the Key Project of State Key Laboratory of Rail Traffic Control and Safety under Crants No. RCS2008ZZ006, No. RCS2011ZZ008 the Program for Changjiang Scholars and Innovative Research Team in University under Crant No. IRT0949 the Project of State Key kab. of Rail Traffic Control and Safety under C~ants No. RCS2008ZT005, No. RCS2010ZT012 the Fundamental Research Funds for the Central Universities under Crants No. 2010JBZ（~8, No. 2011YJS010.
文摘The non-stationary behavior, caused by the train rmverrent, is the rmin factor for the variation of high speed railway channel. To measure the tirce-variant effect, the parameter of stationarity interval, in which the channel keeps constant or has no great change, is adopted based on Zhengzfiou- Xi＇an （Zhengxi） passenger dedicated line measurement with different train speeds. The stationarity interval is calculated through the definition of Local Region of Stationarity （LRS） under three train ve- locities. Furthermore, the time non-stationary characteristic of high speed pared with five standard railway channel is corn- Multiple-Input MultipleOutput （MIMO） channel models, i.e. Spatial Channel Model （SCM）, extended version of SCM （SCME）, Wireless World Initiative New Radio Phase II （WINNERII）, International Mobile Teleconmnications-Advanced （IMT-Advanced） and WiMAX models which contain the high speed moving scenario. The stationarity interval of real channel is 9 ms in 80% of the cases, which is shorter than those of the standard models. Hence the real channel of high speed railway changes more rapidly. The stationarity intervals of standard models are different due to different modeling methods and scenario def- initions. And the compared results are instructive for wireless system design in high speed railway.
