| Summary: | New generation wireless networks are designed not only to carry voice but also to support data-intensive and multimedia applications. Broadband wireless networks offer high bandwidth necessary to support these applications. However, without proper resource allocation schemes, increased bandwidth is not sufficient to meet diverse application quality of service (QoS) requirements. In designing or deploying a resource allocation scheme, it is crucial to understand the inter-relationship of the resource allocation scheme and important system parameters with resulting QoS performance. Analytical models provide an opportunity to derive these relationships in an accurate and readily verifiable way.
In this thesis, we develop novel analytical models for radio resource allocation schemes in emerging broadband wireless access networks. These models are then adopted for in-depth analysis of QoS performance of the modeled schemes and in devising new solutions based on the models to either improve upon or complement those schemes. Our work primarily deals with Medium Access Control layer; however, in most of our contributions, we also consider cross-layer issues. First, we develop a queueing model for a downlink packet scheduling policy in IEEE 802.16e mobile broadband systems and propose a resource allocation framework based on this model. Compared to existing schemes, proposed framework offers a simple yet more effective way to provide QoS to a heterogeneous mix of applications. Second, we develop a cross-layer model for a prominent multiuser scheduling scheme in multi-antenna-based broadband cellular systems. It captures cross-layer effects of important parameters of the multi-antenna physical layer. The model output is shown to have important applications in QoS provisioning. Next, we perform queueing analysis of controlled channel access mechanism in IEEE 802.11e-based Wireless Local Area Networks. Using the insight gained, we propose a novel channel access scheduling mechanism that achieves very robust performance in meeting QoS guarantees. Finally, we focus on a promising new technology called Cognitive Radio (CR), which can greatly improve spectrum utilization in next generation broadband systems. We develop a queueing model to analyze the performance of an opportunistic spectrum access mechanism in CR networks. The model has important applications including cross-layer analysis and admission control in CR-based broadband networks.
|
|---|
