ESBA: hybrid Energy-Saving video Bitrate Adaptation algorithm to deliver videos with high Quality of Experience and energy-efficiency for mobile users
Energy, HAS, QoE
The number of mobile devices that use video streaming applications has been steadily rising year after year. Platforms responsible for providing multimedia service face great challenges in delivering high-quality content for mobile users due to frequent disconnections, often caused by user movements and heavily energy-dependence of mobile devices. A video adaptation approach with Quality of Experience (QoE) and Energy-saving support is a key issue to mitigate these problems, enhancing user QoE, as well as reducing the energy consumption in mobile devices. In this master thesis, we propose a hybrid Energy-Saving video Bitrate Adaptation algorithm (ESBA) to deliver videos with high QoE and energy-efficiency for mobile users. In addition, we consider an Artificial Neural Network (ANN) approach for individual network throughput prediction. Simulation results show the efficiency of ESBA compared to existing adaptation video bitrate algorithms, reducing the number and duration of player stalls, as well as saving energy of mobile devices. Moreover, observing results, we notice that the ANN approach overcomes current throughput prediction approaches present in literature in specific scenarios, permitting the adaptation algorithm to respond more efficiently to network changes.