| With the popularity of mobile Internet,wireless live video service is developing rapidly.In contrast to video-on-demand services,live content is generated in real time and has higher requirements for low latency.As an effective way to cope with the time-varying characteristics of wireless channels,adaptive streaming technology supports live users to dynamically adjust the video bitrate,which has received widespread attention.Considering that a large number of users often watch the same content simultaneously in live scenario,it is suitable to apply wireless multicast technology to share transmission resources,and it is important to study how to effectively combine it with adaptive streaming technology to improve system performance.Meanwhile,optimizing the user experience in a typical unicast scenario is also a key challenge as it is difficult to meet the multicast application conditions if the content requested by different users varies.Therefore,for the above two live scenarios of multicast and unicast,this thesis investigates the multicast transmission optimization scheme and adaptive playback strategy,respectively.Firstly,for the public transmission scenario in wireless live video service,a real-time video multicast system architecture is proposed in this thesis.A network-assisted real-time video multicast transmission optimization scheme based on network assistance is designed under this architecture jointly considering user grouping,resource allocation and bitrate selection problems.First,a K-means++-based user grouping method is applied to determine the number of groups and user grouping results.Second,the joint resource allocation and bitrate selection algorithms for different time scales are proposed based on the Lyapunov optimization method with the fixed user grouping information,and then the theoretical lower bound of the algorithm performance is given.Compared with other benchmark schemes,the proposed multicast scheme in this thesis achieves at least 14%performance improvement in user experience.Secondly,for the typical unicast scenario in wireless live video service,a latency control mechanism is introduced in this thesis in addition to the bitrate selection.The client buffer model is reconstructed with the characteristics of the live video service.A live adaptive playback strategy is designed with bitrate selection and latency control mechanism as the core.The proposed algorithm considers the before-and-after correlation of video playback,and jointly optimizes the bitrate selection and latency control strategies based on a receding horizon control method with the objective of maximizing long-time QoE.Simulation results show that the proposed adaptive playback strategy improves the average QoE by at least 9.1%compared with existing algorithms and achieves a better tradeoff between different performance metrics. |
PDF Full Text Request