Increasing real-time video streaming performance over Wi-Fi networks

This project focuses on improving the performance of the real-time video streaming over Wi-Fi networks, in which losing packets is a serious concern. State-of-the-art methods try to improve the robustness of AL-FEC mechanisms by providing unequal protection to the packets. But such approaches are complex and increase even more the delays. In contrast, the proposed research will study how to dynamically select the packet size and the period of AL-FEC redundant data (number of rows and columns in AL-FEC), crucial parameters for real-time applications. The research will study how to adjust the packet size dynamically based on estimated communication conditions (e.g. the number of packets received without error versus damaged ones, etc.) by using machine learning approaches. Appropriate selection of packet size will optimize the effective throughput as the network conditions change. TO BE CONT'D

Mona Naghashi
Superviseur universitaire: 
Stéphane Coulombe