Abstract

Video streaming has become the dominant traffic type on cellular networks. It continues to present challenges, due to the variability of network performance, increased demand, and greater quality expectations from users. In this seminar, I will focus on our recent work in the iViD project to improve video streaming over wireless networks. This will include using machine learning for throughout prediction and how to employ this effectively in video players. I will also describe our published 4G/5G/video datasets. The approach will be to explain the rationale for the research, the evolution of our solutions, choice of methodologies, and remaining challenges. I will also briefly introduce CONNECT , which is the Irish research centre for future networks and communications, within which much of my work is conducted.