Improving QoE via Context Prediction: A Case Study of Using WiFi Radiomaps to Predict Network Disconnection

Abstract

This paper proposes a novel way to improve the user Quality of Experience (QoE) by monitoring and predicting their context. The method builds on a fingerprint-based indoor positioning system, which monitors the user's position and uses that to predict the quality of the network connection. Successfully predicting the quality of the network connection allows applications which are sensitive to network fluctuation such as video and audio streaming apps to optimize their buffering strategy, thus improving the overall QoE perceived by the end users. Our approach is demonstrated in the context of a case study-based evaluation, using a blend of real and simulated data.