Dynamic resource allocation for uplink packet access in high speed cellular wireless networks

该项目的重点是移动的、无处不在的无线网络的高速上行分组接入（HSUPA）和1xEV-DO。这些技术旨在为蜂窝接入网络的上行链路密集型数据和多媒体服务实现宽带连接。用于用户调度、重传控制和接纳控制的资源分配算法并没有作为这些标准的一部分被具体定义。这促使我们的目标，设计有效的资源分配算法，提高最终用户的QoS和优化的上行链路网络吞吐量。我们的具体目标包括：（i）开发实用的实验为基础的移动性和业务模型在上行链路无线网络。(ii)优化用户调度，媒体访问和admssion控制在个别层和跨层。(iii)为音频和视频上行链路传输设计高效的可伸缩MPEG 4传输方案。(iv)将开发的算法和软件整合到复杂的HSUPA和1xEV-DO测试平台中。上述目标导致我们的长期愿景，推进动态资源分配的理论和实践前沿，使HSUPA和1xEV-DO系统无缝地适应特定的应用和流量类型。我们的研究广泛使用了复杂的分析工具，如超模块化（用于分析最优调度策略的结构），弱收敛（用于分析系统的重流量限制）和随机平均理论（用于分析网络的快速和慢速动态以构建易于处理的模型）。这些工具产生大量的洞察上行链路网络的行为。该项目将涉及与四家公司的实质性合作，这些公司横跨整个无线电信领域：Sierra Wireless（设备制造商）、Telus（服务提供商）、Dyaptive Systems（诊断测试）和MIMOW Technology（多媒体）。

This project focuses on High Speed Uplink Packet Access (HSUPA) and 1xEV-DO for mobile, ubiquitouswireless networks. These technologies aim to achieve broadband connectivity inuplink-intensive data and multimedia services for cellular access networks. Resource allocation algorithms foruser scheduling, retransmisison control and admission control are not specifically defined as part of thesestandards. This motivates our aim of devising efficient resource allocation algorithms that enhance end-userQoS and optimize the uplink network throughput. Our specific goals include: (i) Developing practicalexperimental-based mobility and traffic models in uplink wireless networks. (ii) Optimizing user scheduling,medium access and admssion control at individual layers and across layers. (iii) Devising efficient scalableMPEG4 transmission schemes for audio and video uplink transmission. (iv) Consolidating the algorithms andsoftware developed into a sophisticated HSUPA and 1xEV-DO testbed. The above objectives lead to our longterm vision of advancing the theoretical and practical frontiers of dynamic resource allocation so that HSUPAand 1xEV-DO systems seamlessly adapt to specific applications and traffic types. Our research makesextensive use of sophisticated analytical tools such as supermodularity (for analysing the structure of theoptimal scheduling policy), weak convergence (to analyse the heavy traffic limit of the system), and stochasticaveraging theory (for decouping fast and slow dynamics of the network to construct tractable models). Thesetools yield substantial insight into the behaviour of the uplink network. The project will involve substantialcollaboration with four companies than span the entire range of wireless telecommunications: Sierra Wireless(Equipment manufacturer), Telus (Service provider), Dyaptive Systems (Diagnostic testing), and MIMOWTechnology (multimedia).