QoE-aware Dynamic Participant Selection of Real-time Video Transmission on Smartphones

摘要

Smartphone-based one-tomany real-time video transmission gains its popularity due to the development of smartphones. However, when multiple clients apply for sharing a same video, current systems fail to provide the best service with limited network resources and hardware restrictions, leading to low video quality and poor network performance, which greatly damages the Quality of Experience (QoE) of clients. In this paper, a QoE-aware dynamic participant selection algorithm of real-time video transmission is proposed to optimize the QoE of clients by figuring out the most appropriate participant group and transmitting the highest quality videos. A Multi-objective Programming (MOP) problem is put forward and broke down into three sub optimization functions, which are converted into knapsack problem, dynamic programming and linear programming then solved with layered method, decision tree and evaluation method. Experiments show that the algorithm has higher performance than systems that select participants randomly and the video transmitted is fluent and clear with good client experience. Therefore our proposed algorithm is suitable for real-time one-to-many video transmission.