Research And Implementation Of QoE-driven Adaptive Transmission Control Mechanism Of Tile-based Panoramic Video

With the continuous advancement of Internet technology and the rapid popularization of 5G communication,panoramic video has attracted widespread attention.Compared with traditional video,the viewport(user viewing area)of panoramic video is only a part of the entire video.If the entire panoramic video area is transmitted during user viewing,a large amount of bandwidth resources will be wasted,and the video quality in the viewport area will be reduced.The tiled panoramic video adaptive transmission technology transmits the viewport in tiles,fine-grainedly adjusts the bit rate of different spatial positions of the panoramic video,and improves the video quality of the user’s viewport area.However,in the case of network jitter and changing user viewports,how to allocate bit rate and video resource transmission for tiled panoramic video will affect the user’s quality of experience(QoE),but the existing methods do not consider how to prevent video interruption and video quality degradation caused by network prediction and viewport prediction failures during the allocation process of bit rate,and do not consider the high delay of video resources in tiles transmission and video black tiles caused by network changes during transmission.The QoE of panoramic video cannot be guaranteed.This paper mainly studies how to ensure the QoE of panoramic video in the process of bit rate allocation and resource transmission under the scene of network jitter and user viewport change.First of all,this paper proposes a buffer-based rate allocation mechanism with fault-tolerant control to solve the problem of panoramic video QoE degradation when network bandwidth prediction fails and user viewport prediction fails.The mechanism defines the QoE evaluation model,and uses the code rate adjustment function and buffer replacement method to determine the total code rate of the request tiles and the set of video tiles that need to be rerequested,so as to ensure the QoE of the panoramic video when the prediction fails.Secondly,in order to reduce the request delay of the request tiles and reduce the occurrence of video black tiles and video interruptions,we propose a bit rate feedback mechanism based on the HTTP/2 protocol,which prioritizes the request tiles and utilizes the server-active push and parallel transfer features of HTTP/2,real-time feedback of network conditions during transmission and corresponding adjustment of video bit rate to adapt to network changes.Finally,this paper builds a QoE-driven tile-based panoramic adaptive transmission control system,which realizes the fault-tolerant mechanism of panoramic video bit rate distribution based on buffer control and the bit rate feedback mechanism based on HTTP/2 protocol,and tests the system function and mechanism performance.The experimental results show that the mechanism proposed in this paper can effectively improve the panoramic video QoE in the scene of network jitter and user viewport change.