A New Variable Bit Rate (VBR) Video Traffic Model Based on Fuzzy Systems Implemented Using Generalized Regression Neural Networks (GRNN)

Abstract

Variable Bit Rate (VBR) video traffic modeling is one of the most important challenges in setting up the communication networks. It is applicable to the efficient designing and using of the networks. There are some solutions to the problem of VBR Video traffic modeling. These solutions are divided in to two categories: statistical modeling and deterministic modeling. In this paper, we propose a new deterministic model based on the nonlinear mapping characteristic of fuzzy systems, implemented using a Generalized Regression Neural Network (GRNN). This model is tested with prediction of next 5, 10 and 20 frames of I, P and B frames of the desired movie, using 100 current frames. The model generated data is examined by the matching of autocorrelation function and histogram. This model can regenerate the ACF and histogram of video traces, in order of 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-3</sup> in error.