CAREER: Efficient Resource Management and Multi-Access Protocols for Bursty Traffic Over Wireless Networks: A Cross-Layer Design Approach

近年来，对无处不在的信息访问的需求急剧增长。然而，目前的无线网络还远远不能满足这一需求。无线系统设计中的一个核心问题是如何在无线链路上有效地传输突发多媒体业务。预计开发利用通信通道和上层协议层之间相互作用的网络级解决方案将产生显著的性能增益。跨多层的优化设计打开了一个新的有前途的领域，许多设计问题尚未解决。因此，发展有助于推动这一领域取得重大进展并导致革命性突破的理论和方法至关重要。在本项目中，PI建议采用跨层设计方法来设计一套资源分配和多址方案，旨在建立一个在衰落信道上传输突发业务的综合框架。提出的研究主要围绕两个方面展开：1)CDMA突发流量：该研究的一个关键目标是获得限制干扰系统中突发流量传输跨层优化的首要原则；2)突发流量的“机会主义”访问控制：这一重点旨在深入了解如何利用交通信息在机会主义通信系统（基本上是配备机会主义调度的TDMA系统）中进行新型访问控制。在整个过程中遇到的一个常见问题是“利用”（而不是“对抗”）流量突发和信道变化。1) CDMA网络中的突发业务传输：基于PI最近发现的多址干扰（MAI）是远程依赖的，PI将对MAI的远程依赖进行全面研究，确定预测MAI结构，并利用MAI结构开发有效的基于测量的干扰预测。研究了交通突发性、衰落和反馈延迟的影响；B.利用干扰预测探索有效的资源分配和访问控制。2)机会通信系统中的突发流量传输：PI将：a.研究机会通信系统的交通辅助准入控制；B.为流媒体设计创新的机会调度，以利用嵌入在信道变化和流量突发中的多用户多样性增益。

Recent years have witnessed a tremendous growth in the demand for ubiquitous information access. Current wireless networks, however, are still far from meeting this demand. A central problem in the design of wireless systems is how to efficiently transmit bursty multimedia traffic over wireless links. It is expected that developing network-level solutions that take advantage of the interplay between the communication channel and the upper protocol layers would yield significant performance gains. Optimal design across multiple layers opens a new promising area with many design issues unresolved. It is therefore of vital importance to develop theory and methodology that help propel significant advances and lead to revolutionary breakthroughs in this area.In this project, the PI proposes to take a cross-layer design approach to devising a suite of resource allocation and multi-access schemes, aiming to establish a comprehensive framework for transmitting bursty traffic over fading channels. The proposed research is centered around two areas: 1) bursty traffic over CDMA: a key goal of this thrust is to obtain overriding principles for cross-layer optimization of bursty traffic transmissions in interference-limited systems; and 2) "opportunistic" access control for bursty traffic: this thrust aims at a deep understanding of how to exploit traffic information for novel access control in opportunistic communication systems (which are basically TDMA systems equipped with opportunistic scheduling). A common thread encountered throughout is to "exploit" (rather than "combat") traffic burstiness and channel variation. The two thrusts are outlined as follows:1) Bursty traffic transmission in CDMA networks: Building on the PI's recent finding that the multi-access interference (MAI) is long-range dependent, the PI will a. conduct a comprehensive study on the MAI long-range dependence and identify the predictive MAI structure, and exploit the MAI structure to develop efficient measurement-based interference prediction. The impact of traffic burstiness, fading, and feedback delay will be examined; b. utilize interference prediction to explore efficient resource allocation and access control.2) Bursty traffic transmission in opportunistic communication systems: The PI will: a. investigate traffic-aided admission control for opportunistic communication systems; b. devise innovative opportunistic scheduling for streaming multimedia to exploit multi-user diversity gain embedded in both channel variation and traffic burstiness.