Energy optimized mobile video streaming with user behavior

Abstract

With the rapid development of mobile devices such as mobile phones, the internet traffic generated by mobile streaming media quickly surpassed all previous online entertainment applications. In order to reduce the energy consumption of mobile device users watching videos, we introduce an innovative Energy-Optimized Mobile Video Streaming Transmission System (EVSU) rooted in user behavior analysis. Leveraging a robust language model, EVSU scrutinizes video subtitle data, synthesizes content summaries, and predicts user viewing patterns. Employing an adaptive bitrate algorithm grounded in deep reinforcement learning, EVSU factors in users' quality of experience (QoE), data efficiency, and energy consumption to dynamically select optimal video bitrates. Experimental results, using the popular documentary Planet Earth, showcase EVSU's remarkable ability to significantly reduce data transfer and energy consumption compared to existing algorithms, all while upholding a high user-perceived quality.