首页> 外文学位 >Wireless networks: Model and optimization based approaches to clock synchronization, random access MAC and video streaming.
【24h】

Wireless networks: Model and optimization based approaches to clock synchronization, random access MAC and video streaming.

机译:无线网络:基于模型和优化的时钟同步,随机访问MAC和视频流的方法。

获取原文
获取原文并翻译 | 示例

摘要

We, via a model and optimization-based approach, address three issues related to wireless networks: clock synchronization, medium access control (MAC) and scalable video streaming.;In Chapter 2 we develop, study and simulate a new model-based distributed network clock synchronization protocol. We use the distributed asynchronous algorithm to obtain nodal offset estimates from relative offset estimates. We combine our findings into developing a new protocol for clock synchronization, namely the Model-Based Clock Synchronization Protocol (MBCSP). We present a comparative simulation study of its performance versus the leading scheme by Solis et al. (2006); the results show that MBCSP performs better in terms of skew, offset and delay estimation. Finally, we have performed trace-driven simulation based on time-stamps obtained from Berkeley motes. Our scheme outperforms that of Solis et al. by 45%, where we used the accuracy in predicting the receipt time-stamp at the sender as the clock synchronization metric.;In Chapter 3, we study random access based MAC in the framework of network utility maximization (NUM). There has been much recent interest in protocol design for wireless networks based on maximizing a network utility function. A significant advance is the observation that a decomposition of the Lagrangian suggests an approach where transmissions are scheduled to minimize back-pressure. However, a satisfactory MAC protocol that can realize such a scheduling algorithm is notably missing, and we develop one potential scheme. We present a candidate random access MAC protocol that extends an existing algorithm by Gupta and Stolyar (2006) in calculating the access probabilities. We also consider the online adaptation of access probabilities using local information about queue lengths and active links. We provide OPNET simulation results to compare the performance of our scheme with the leading schemes. We estimate the capacity region of our scheme by simulation for various topologies and multiple flows. Our simulation studies indicate that our extension in conjunction with an implementation of back-pressure significantly outperforms the slotted-time algorithm of Gupta and Stolyar (2006).;In Chapter 4, we present performance bounds for random access based MAC using carrier-sense multiple access (CSMA). In recent work, it was shown that a distributed CSMA-based MAC protocol is throughput-optimal which, in turn, implies that the class of controlled distributed random access MAC protocols can support the entire capacity region. It is challenging to study the performance of such schemes in terms of mean delays and compare it with some known results on the performance of centralized scheduling. We modify the model of Jiang and Walrand (2008) to obtain Markov chain models that incorporate the queue lengths as well as the information about the independent set, for single-hop networks. We show that the delay of the new models yields an upper bound on the delay of the original models. We derive upper and lower bounds on the mean total delay at the steady-state, and show that these bounds coincide with those for max-weight scheduling. Finally, we develop a method of deriving upper and lower bounds for random-access schemes by using linear programs (LPs). We present an optimization program for minimizing the upper bounds.;In Chapter 5, we consider multihomed scalable video streaming systems where each video is concurrently transmitted over several access networks to a client. The problem is to determine which video packets of a video stream to transmit, and associate each video packet with an access network, so that the video quality at the client is maximized under measured network conditions. We present a network model and a video distortion model to capture the network conditions and video distortion characteristics, respectively. We develop a mathematical formulation to find the streaming strategy for maximizing the average video quality at the client. (Abstract shortened by UMI.)
机译:我们通过基于模型和优化的方法,解决了与无线网络相关的三个问题:时钟同步,媒体访问控制(MAC)和可伸缩视频流。;在第二章中,我们开发,研究和模拟了一种基于模型的新分布式网络时钟同步协议。我们使用分布式异步算法从相对偏移量估计中获得节点偏移量估计。我们将发现结合起来,开发出一种新的时钟同步协议,即基于模型的时钟同步协议(MBCSP)。我们将其性能与Solis等人的领先方案进行对比仿真研究。 (2006);结果表明,MBCSP在偏斜,偏移和延迟估计方面表现更好。最后,我们基于从伯克利微粒获取的时间戳执行了跟踪驱动的仿真。我们的方案优于Solis等人的方案。 45%,其中我们将预测发送方的接收时间戳的准确性用作时钟同步度量。在第3章中,我们在网络效用最大化(NUM)框架下研究了基于随机访问的MAC。基于最大化网络实用程序功能的用于无线网络的协议设计已经引起了很多兴趣。一个重要的进步是观察到拉格朗日分解法提出了一种安排变速箱以最小化背压的方法。然而,明显缺少能够实现这种调度算法的令人满意的MAC协议,因此我们开发了一种潜在的方案。我们提出了一种候选随机访问MAC协议,该协议扩展了Gupta和Stolyar(2006)在计算访问概率时的现有算法。我们还考虑使用有关队列长度和活动链接的本地信息对访问概率进行在线调整。我们提供OPNET仿真结果,以比较我们的方案与领先方案的性能。我们通过仿真各种拓扑和多种流量来估计我们方案的容量区域。我们的仿真研究表明,我们的扩展结合背压的实现明显优于Gupta和Stolyar(2006)的时隙算法。在第4章中,我们给出了使用载波感知倍数的基于随机访问的MAC的性能范围访问(CSMA)。在最近的工作中,已经表明基于分布式CSMA的MAC协议是吞吐量最佳的,这反过来意味着受控的分布式随机访问MAC协议类别可以支持整个容量区域。从平均时延的角度研究此类方案的性能,并将其与集中式调度性能的一些已知结果进行比较是具有挑战性的。对于单跳网络,我们修改了Jiang和Walrand(2008)的模型,以获得包含队列长度以及有关独立集的信息的Markov链模型。我们表明,新模型的延迟产生了原始模型延迟的上限。我们推导出稳态下平均总延迟的上限和下限,并表明这些界限与最大权重调度的界限一致。最后,我们开发了一种使用线性程序(LP)推导随机访问方案的上限和下限的方法。我们提出了一个最小化上限的优化程序。在第5章中,我们考虑了多宿主可伸缩视频流系统,其中每个视频通过多个访问网络同时传输到客户端。问题在于确定要传输的视频流中的哪些视频数据包,并将每个视频数据包与接入网络相关联,以便在测量的网络条件下使客户端的视频质量最大化。我们提出了一个网络模型和一个视频失真模型来分别捕获网络状况和视频失真特征。我们开发了一种数学公式来查找流媒体策略,以最大程度地提高客户的平均视频质量。 (摘要由UMI缩短。)

著录项

  • 作者

    Freris, Nikolaos Markos.;

  • 作者单位

    University of Illinois at Urbana-Champaign.;

  • 授予单位 University of Illinois at Urbana-Champaign.;
  • 学科 Engineering Electronics and Electrical.
  • 学位 Ph.D.
  • 年度 2010
  • 页码 101 p.
  • 总页数 101
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号