隐含自回归视频业务模型

Embedded autoregressive video source model

在无线移动通信领域中 ,为提高接入控制中对速率预测的精度 ,作者提出了一类新的视频信源模型——隐含自回归模型。该模型利用伽马分布随机变量的立方根服从正态分布这一特性成功地仿真了视频业务。仿真实验说明无论是分布特性、复用特性还是码率变化轨迹 ,新模型均优于传统的视频信源模型。因此 ,隐含自回归视频信源模型有效地解决了研究视频业务接入无线异步传输模式 (WATM)网特性时缺乏大量实际信源的难题 ,同时为设计高效的接入算法提供了思路。

The conflicts between limited wireless resource and emerging video services can be resolved using a novel embedded autoregressive (EAR) video source model, which was developed from the theorem that the cubic root of gamma distributed variable has a normal distribution. Simulation results with the new model perform better than conventional methods for distribution, multiplexing property and sample path. Therefore, EAR model will not only provide abundant video sources for the study of access property of multimedia services in wireless ATM network, but also will provide new ideas for high efficiency access methods.