Quality of Experience and Adaptation Techniques for Multimedia Communications

The widespread use of multimedia services on the World Wide Web and the advances in end-user portable devices have recently increased the user demands for better quality. Moreover, providing these services seamlessly and ubiquitously on wireless networks and with user mobility poses hard challenges. To meet these challenges and fulfill the end-user requirements, suitable strategies need to be adopted at both application level and network level. At the application level rate and quality have to be adapted to time-varying bandwidth limitations, whereas on the network side a mechanism for efficient use of the network resources has to be implemented, to provide a better end-user Quality of Experience (QoE) through better Quality of Service (QoS). The work in this thesis addresses these issues by first investigating multi-stream rate adaptation techniques for Scalable Video Coding (SVC) applications aimed at a fair provision of QoE to end-users. Rate Distortion (R-D) models for real-time and non real-time video streaming have been proposed and a rate adaptation technique is also developed to minimize with fairness the distortion of multiple videos with difference complexities. To provide resiliency against errors, the effect of Unequal Error protection (UXP) based on Reed Solomon (RS) encoding with erasure correction has been also included in the proposed R-D modelling. Moreover, to improve the support of QoE at the network level for multimedia applications sensitive to delays, jitters and packet drops, a technique to prioritise different traffic flows using specific QoS classes within an intermediate DiffServ network integrated with a WiMAX access system is investigated. Simulations were performed to test the network under different congestion scenarios.