论文部分内容阅读
分析了ＡＴＭ 复用器中的分组语音图像流。为研究多信源、大流量条件下的流量特性，我们将语音源近似为马尔柯夫泊松过程ＭＭＰＰ（２）．２个泊松过程的平均到达率由生灭链得出，转移率则通过极限定理，由高斯过程的矩匹配ＭＭＰＰ（２）相应参数而获得。同时，我们运用流体流法来解决复用缓冲性能。同样的方法适用于图像源及语音图像合成源的情况。
The packet voice stream in ATM multiplexer is analyzed. To study the flow characteristics under multi-source and large flow conditions, we approximate the speech source to the Markov-Poisson process MMPP (2). The average arrival rate of two Poisson processes is derived from the birth-death chain, and the transfer rate is obtained by the limit theorem from the moment of the Gaussian process matching the corresponding parameters of MMPP (2). At the same time, we use fluid flow method to solve the reuse buffer performance. The same method applies to image source and speech image synthesis source.