ADAPTIVE WIRELESS NETWORKING FOR VIDEO STREAMING
Dr Chan Mun Choon, Associate Professor, School of Computing
COM2 Level 2
MR3, COM2-02-26
Abstract:
Although wireless technologies have evolved significantly over the past decades, the wireless bandwidth, however, is still insufficient to support the fast-growing mobile video streaming traffic. In this thesis, we work towards designing effective solutions in improving the video streaming quality and the network scalability.
First, we investigate mixed resolutions tiling video where tiles in a video frame can come from different resolution streams. Applying this scheme to adaptive wireless multicast, we have the following optimal adaptive multicast allocation problem: given the subset of tiles that each user requested, the link rate of each user, and the available time slots, at which resolution and at which link rate each tile should be sent, to maximize the overall video quality by all users. By applying dynamic programming, we design an efficient algorithm to optimally solve the problem above.
Second, we explore the general multi-sessions multicast allocation problem, where multiple access points are deployed to improve the network capacity. We present a joint user and rate allocation scheme for video multicast over multiple access points. This scheme intelligently determines user to access point association, the video resolution version to be delivered for each session, and the transmission link rate for each video version.
Last, we study the heterogeneous topologies that are present when neighboring access points are operating on the same channel. The saturated traffic introduced by video streaming could lead to severe unfairness in presence of these topologies. We build an analytical model to characterize the network performance. Based on the model, we suggest an adaptive contention window tuning mechanism which is able to effectively remedy the unfair issues.