电信科学 ›› 2020, Vol. 36 ›› Issue (12): 65-76.doi: 10.11959/j.issn.1000-0801.2020308

• 研究与开发 • 上一篇    下一篇

基于A3C的无线异构网络自适应视频流传输控制方法

罗志强,王伟,朱晓荣   

  1. 南京邮电大学通信与信息工程学院,江苏 南京 210003
  • 修回日期:2020-12-02 出版日期:2020-12-20 发布日期:2020-12-23
  • 作者简介:罗志强(1997- ),男,南京邮电大学通信与信息工程学院在读,主要研究方向为无线异构网络视频流自适应传输|王伟(1997- ),男,南京邮电大学通信与信息工程学院硕士生,主要研究方向为无线异构网络视频流自适应传输|朱晓荣(1977- ),女,博士,南京邮电大学通信与信息工程学院教授、博士生导师,主要研究方向为下一代无线网络、异构网络

An adaptive video stream transmission control method for wireless heterogeneous networks based on A3C

Zhiqiang LUO,Wei WANG,Xiaorong ZHU   

  1. School of Communication and Information Engineering,Nanjing University of Posts and Telecommunications,Nanjing 210003,China
  • Revised:2020-12-02 Online:2020-12-20 Published:2020-12-23

摘要:

比特率自适应(ABR)算法已经成为视频传输中研究的热点之一。然而,由于5G无线异构网络具有信道带宽波动大、不同网络间差异明显等特点,多终端协同的自适应视频流传输面临着巨大挑战。提出了一种基于深度强化学习的自适应视频流传输控制方法。首先,建立了视频流动态规划模型,对传输码率以及分流策略进行联合优化。由于该优化问题的求解依赖于精确的信道估计,这在信道状态动态变化的网络中很难实现。因此,将动态规划问题改进为强化学习任务,并采用A3C算法,动态决策视频码率和分流策略。最后,根据实测的网络数据进行仿真,与传统的优化方法相比,本文所提的方法较好地提高了用户QoE。

关键词: 无线异构网络, A3C, 码率自适应, 多终端协同, QoE

Abstract:

The adaptive bit rate (ABR) algorithm has become the focus research in video transmission.However,due to the characteristics of 5G wireless heterogeneous networks,such as large fluctuation of channel bandwidth and obvious differences between different networks,the adaptive video stream transmission with multi-terminal cooperation was faced with great challenges.An adaptive video stream transmission control method based on deep reinforcement learning was proposed.First of all,a video stream dynamic programming model was established to jointly optimize the transmission rate and diversion strategy.Since the solution of this optimization problem depended on accurate channel estimation,dynamically changing channel state was difficult to achieve.Therefore,the dynamic programming problem was improved to reinforcement learning task,and the A3C algorithm was used to dynamically determine the video bit rate and diversion strategy.Finally,the simulation was carried out according to the measured network data,and compared with the traditional optimization method,the method proposed better improved the user QoE.

Key words: wireless heterogeneous network, A3C, rate adaptation, multi-terminal cooperation, QoE

中图分类号: 

No Suggested Reading articles found!