多用户场景下基于MPVRT模型的 体育视频传输模型研究

VR Sports Video Transmission Model in Multi-user Scenarios

针对多用户场景下的体育视频传输中,虚拟现实(virtual reality,VR)技术易出现带宽资源限制、延时过高等问题,实验提出一种结合多径协同算法的VR体育视频传输模型。首先对VR视频传输系统的结构进行了阐释,在此基础上引入多路径协议来提高网络的吞吐率和利用率。此外,实验还利用强化学习算法来改善多路径协议的性能。最终实验结果显示,MPVRT模型的平均吞吐量可高出同类型算法8.64Mbps,其时延低于同类模型0.33S。由此,MPVRT模型具有更高的网络吞吐量和更少的应用时延,能够给用户带来良好的体验。

Aimed at the problems of bandwidth resource limitation and high delay in virtual reality(VR)technology in multi-user scenarios,a VR sports video transmission model combined with multi-path collaborative algorithm was proposed in the experiment,that is the multi-path virtual reality sports video transmission(MPVRT)based on multi-path collaboration.Firstly,the structure of VR video transmission system was explained,and then multipath protocol was introduced to enhance the throughput and utilization rate of the network.In addition,reinforcement learning algorithm was applied to improve the performance of multipath protocol.The final experimental results showed that the average throughput of MPVRT model was 8.64 Mbps higher than that of the same type algorithm,and its delay was 0.33s lower than that of the same type algorithm.Therefore,MPVRT model has higher network throughput and less application delay,which can bring good experience to users.