ATM网的双Gamma自回归视频模型

The Two Gamma Autoregressive Video Model on ATM Networks

本文提出了一种新的ＶＢＲ视频业务模型：三态马尔科夫链下的双一阶自回归Ｇａｍｍａ序列（２ＧＡＲ（１））．该模型的Ｇａｍｍａ分布特性能较好地描述ＶＢＲ业务比特流的分布，由自相关函数决定的模型系数更充分地描述了自相关函数的行为，利用三态马尔科夫链的两个补充过程可以较理想地描述场景切换下的视频业务模型．文中对这一模型进行了较详细的分析，给出了确定模型参数的方法，并得出了一些重要结论．

In this paper,we propose a new VBR video model,a double first order autoregressive Gamma sequence with a three state Markov chain (2 GAR(1)).First,its characteristic of Gamma distribution can follow the bit rate of a VBR video source better.Second,the model's coefficients determined by the autocorrelation function can explain the function's behavior better.Third,the modeling of video source is accomplished using the three state Markov chain including the effects of scene change.We analyze the model in detail.Also,we give the methods to determine the parameters,and reach some important conclusions.