CIF: Small: Analog Networking: Distributed Source-Channel Approaches to Delay and Resource Constrained Communications

CIF：小型：模拟网络：解决延迟和资源受限通信的分布式源通道方法

• 批准号: 1118075
• 负责人: Kenneth Rose
• 金额: $ 30.21万
• 依托单位: University of California-Santa Barbara
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2011
• 资助国家: 美国
• 起止时间: 2011-10-01 至 2015-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1118075&HistoricalAwards=false
• 关键词: CIF; Small; Analog; Networking; Distributed

Severe delay and resource (energy, bandwidth) constraints are typical in emerging communications and networking applications within a variety of scientific and engineering disciplines. This is particularly the case in highly constrained sensor networks. Such applications strongly motivate foregoing the convenience of digital communication and revisiting the potential benefits, and even necessity of analog communication, albeit in the modern context of distributed source-channel coding and routing. Major challenges emerge, including the derivation of a theoretical foundation for analog networking from estimation and information theoretic principles, specializing to a variety of subproblems and networking scenarios involving distributed source-channel coding, and developing effective practical algorithmic approaches, as well as a formidable optimization challenge due to the high complexity of the overall objective function which is literally riddled with poor local optima.The general analog networking problem is formulated in terms of analog mappings, between source and channel variables at source nodes (encoders), between received and transmitted variables at intermediate network nodes, and between received and reconstruction variables at sink nodes (decoders). Research is pursued along several specific lines of investigation, including derivation of delay-constrained performance bounds and optimality conditions, optimal and low complexity algorithm design, global optimization techniques, analog mappings for distributed source-channel coding, analog multiple descriptions coding, extensions to sources and channels with memory, and optimal routing in an analog network. The effort builds on a body of preliminary work and results, including the necessary conditions for optimality of such mappings in simple settings, iterative algorithms to find locally optimal mappings, as well as a powerful non- convex optimization tool (deterministic annealing), which has been successfully employed in related optimization problems.

严重的延迟和资源（能量、带宽）限制在各种科学和工程学科的新兴通信和网络应用中是典型的。在高度受限的传感器网络中尤其如此。尽管在分布式源信道编码和路由的现代背景下，此类应用强烈地促使人们放弃数字通信的便利性并重新审视模拟通信的潜在好处甚至必要性。出现了重大挑战，包括从估计和信息论原理中推导出模拟网络的理论基础，专门研究涉及分布式源信道编码的各种子问题和网络场景，开发有效的实用算法方法，以及强大的优化由于整体目标函数的高度复杂性而面临挑战，该函数实际上充满了较差的局部最优值。一般的模拟网络问题是根据源节点（编码器）的源变量和通道变量之间、接收变量和传输变量之间的模拟映射来制定的在中间网络节点处，以及在接收器节点（解码器）处接收到的变量和重建变量之间。研究沿着几个具体的研究方向进行，包括延迟约束性能界限和最优条件的推导、最优和低复杂度算法设计、全局优化技术、分布式源通道编码的模拟映射、模拟多重描述编码、源扩展和带有记忆的通道，以及模拟网络中的最佳路由。这项工作建立在一系列初步工作和结果的基础上，包括在简单设置中实现此类映射最优的必要条件、寻找局部最优映射的迭代算法，以及强大的非凸优化工具（确定性退火），该工具具有并成功应用于相关优化问题。