Enhancing Quality of Video Services over Wireless Networks

As development of wireless networking technologies continues, it has been increasingly expected to support content-rich multimedia communications (e.g., video) in such networks. In this dissertation, we study the important problems of enhancing QoS support for video applications in wireless mesh networks and LTE networks.;We begin our investigation with a video performance evaluation study in a wireless mesh test-bed. The goal is to clearly understand the nature of video streaming and its performance in wireless mesh networks. By extensive experiments in various scenarios we analyze the network parameters and settings which impact video performance. Compared with simulation or network-layer statistics based approaches, our investigation is directly focused on video quality in multi-hop scenarios. The experimental results better represent the characteristics of real networks and reveal the interesting aspects of video performance in mesh networks, which may benefit the design of efficient QoS solutions for multimedia services in wireless mesh networks.;Next, we explore the interference issue in wireless mesh networks, and propose an interference-aware admission control and QoS routing scheme, towards enhancing video streaming experience in wireless mesh networks. Interference is characterized using a conflict graph based model. Nodes exchange their flow information periodically and compute their available residual bandwidth based on the local maximal clique constraints. Admission decision is made based on the residual bandwidth at each node. An on-demand routing scheme is implemented to explicitly incorporate the interference model in the route discovery process. By avoiding the ``hot-spots'' with severe interference, it discovers the routes with less interference and enhances video performance significantly.;We also leverage MAC layer retransmission counts to design a combined queuing and routing scheme which achieves better video quality than the existing approaches. A case study is performed focusing on MAC layer retransmission, which helps us better understand the impacting factors on video quality in multi-hop mesh networks. Our scheme is proposed from a cross-layer perspective which leverages the information across network (routing) layer and link (MAC) layer. MAC layer retransmission count is exploited to guide the interface queue management for video packets as well as a metric to construct a retransmission-aware QoS routing scheme. In our approach, the upper layers are aware of the dynamic network status via retransmission count, so timely QoS decision can be made to enhance video quality effectively.;For LTE networks, our focus is radio resource management. We first investigate the important problem of downlink resource allocation in recently enhanced LTE-Advanced systems where a newly added feature, carrier aggregation, provides more flexibility in resource allocation in addition to the existing resource block level packet scheduling. The resource allocation problem is formulated as a complex combinatorial problem with multiple constraints. We decompose this highly complex optimization problem and construct a two-tier resource allocation framework which incorporates dynamic component carrier assignment and novel backlog based scheduling schemes with intelligent link adaptation based on weighted CQI. The proposed schemes offer both better throughput and delay performance as well as user fairness.;Finally, we address the downlink scheduling problem in LTE networks by directly optimizing video quality. Unlike the conventional scheduling rules which exploit network layer metrics, our scheme is directly targeted on optimizing application-layer video quality. Video quality optimized scheduling is formulated as a complex combinatorial optimization problem, and solved by exploiting a GA (genetic algorithm) based metaheuristic approach. The intrinsic strength of population based solution of GA offers a superior advantage for this type of scheduling problems. Simulation results demonstrate the effectiveness of GA based scheduling scheme compared with the well-known M-LWDF scheduling rule and the simulated annealing method. It can offer better video quality at the receiver within the required delay bound.