Business Systems and Knowledge Modelling Lab

Research Project

Multimedia and Networking

Research Goal
This project proposes a intelligent traffic shaping algorithm based on neural networks, which adapts to a network over which streaming video is being transmitted. The purpose of this intelligent shaper being to eradicate all traffic congestion and improve the end-user’s video quality. It possesses the capability to predict, to a very high level of accuracy, a state of congestion based upon the training data collected about the network’s behavior.

A simulation in a controlled environment is illustrated to exhibit the effects of this intelligent traffic-shaping algorithm on the underlying network’s real time packet traffic and the eradication of unwanted abruptions in the streaming video’s quality. Preliminary results of the simulation show that neural networks are a very superior means of modeling real-time traffic and that it can be applied as an appropriate solution to network congestion.

Team Members
Manohar Kaul, Rajiv Khosla and Tasue Mitsukura

Publications
[1] M. Kaul, R. Khosla and Y. Mitsukura, “Intelligent Packet Shaper to Avoid Network Congestion for Improved Streaming Video Quality at Clients,” in IEEE 2003 International Symposium on Computational Intelligence for Robotics and Automation, IEEE Computer Society, Kobe, Japan, July 2003