Measuring and predicting quality of experience of DASH-based video streaming over LTE

Mobile video service becomes more and more popular with commercial deployment of Long Term Evolution (LTE) networks and other wireless access mechanisms. However, perfect LTE coverage with ubiquitous signal quality is not available in reality. Recently, adaptive HTTP video streaming over LTE has been gaining popularity. The quality of video service faces challenges. In this paper, we study the Quality of Experience (QoE) problems of Dynamic Adaptive Streaming over HTTP (DASH) video in LTE networks, building two machine learning models to predict user QoE based on LTE network quality metrics. The first model can be used by an user to estimate user QoE before he/she starts watching a DASH-based video, and the second model can be used by DASH adaptive bitrate control module to adjust its video bitrate in real time. We clearly demonstrate that the most important feature that causes QoE degradation of DASH-based video is Round Time Trip (RTT). Our prediction model and research results are beneficial for improvement and design of DASH adaptive bitrate control algorithm over LTE.