Enabling QoE Learning and Prediction of WebRTC Video Communication in WiFi Networks

Real-time video communication is becoming more and more important in our daily life. WebRTC-based video communication technique has attracted a lot of attentions recently. Its user Quality-of-Experience (QoE), however, faces challenges in wireless networks. This paper studies the problem of accurate prediction of QoE of WebRTC in WiFi networks. We first propose a real-time video QoE metric which is based on the time interval between two consecutively played video frames, and prove it correctly reflects playback freezing and video quality. We then conduct 620 experiments in an indoor WiFi environment to evaluate the correlation between video QoE and wireless network conditions. We final build two machine learning models to predict QoE based on wireless network QoS metrics. The first model can be used by a user to estimate her QoE before she initializes a video call, and the second model is for the system to adjust strategy during a video call. Experimental results show that the models are accurate, with F1 scores above 70%. Our results also clearly demonstrate that current WebRTC’s QoE problem is mainly related to the volatility of RTT. Our QoE evaluation method, results, and prediction models are beneficial for wireless video communication system design.