Optimization of Rateless Code Based Video Multicast
MetadataShow full item record
Multimedia services have become one of the major demands in wireless systems. As a result of growing demands for media services, traffic in wireless networks are increasing. Hence, optimization of multimedia delivery systems to efficiently consume the valuable transmission resources in wireless networks has gained a lot of interest. Raptor codes, with linear encoding and decoding time complexity are one branch of fountain codes (also known as rateless codes) which have found their ways in many recent communication standards as application layer forward error correcting (FEC) codes. Various attempts have been made in order to adapt these codes to wireless channels with their time varying nature. When multimedia delivery is targeted, some other issues such as delay should also be considered. Moreover, in multicast solutions, the system has to address demands of multiple clients. In this thesis, we investigate some optimization scenarios for wireless multimedia multicast systems wherein clients with heterogeneous channels and media quality demands subscribe to a video program. The video program is assumed as a multilayer source with possible spatial, temporal and fidelity layers. The point of optimization under various systems is to provide the clients of different quality constraints with their demanded services while imposing the minimum network/client cost (e.g. delay, power consumption, outage probability) or maximize the provided utility to the clients while considering their heterogeneous capabilities. To achieve these tasks, we study outage probability which serves as a measure to quantify the reliability of a service in a client's side. Packetized rateless multimedia multicast (PRMM) with few optimization criteria regarding the experienced delay in clients are studied and analytical solutions are obtained. A new optimization framework for rateless multimedia multicast is proposed in which, the provided utility to heterogeneous clients are maximized with respect to the clients channel and their quality demands. Application of this optimization in a rateless multimedia multicast system wherein the utility is defined based on perceptual quality experience of clients is also investigated.