CIF: Small: Collaborative Research: Communications in Ultra-Low-Rate Regime: Fundamental Limits, Code Constructions, and Applications

CIF：小型：协作研究：超低速率制度下的通信：基本限制、代码构造和应用

• 批准号: 1910056
• 负责人: Hamed Hassani
• 金额: $ 25万
• 依托单位: University of Pennsylvania
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-10-01 至 2023-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1910056&HistoricalAwards=false
• 关键词: CIF; Small; Collaborative; Research; Communications

Wireless networks have become an integral part of our society by connecting billions of users around the globe. The next generation wireless networks, known as 5G, are expected to deliver orders of magnitude higher data rates to an ever-increasing number of devices. The fundamental limits on how much networks can be pushed towards these advancements are determined by the amount of available wireless bands. To this end, two extremes of wireless communications have become increasingly important, namely, narrowband and wideband. The narrowband communication, using a small band of frequencies, enables connectivity of thousands of devices, such as everyday objects in Internet-of-Things (IoT) networks, to a single cell tower. Wideband communications, using large swaths of frequencies, at extremely high frequencies is the key enabling technology for delivering ultra-high data rates to 5G mobile users. In both of these scenarios, individual bits of information are assigned extremely low power for communication resulting in bit-wise wireless channels with low capacity. This introduces a new array of fundamental problems with respect to communications over low-capacity channels that is the focus of this project. The results of this project will contribute to the foundation of information and coding theory as well as the development of wireless networks that can achieve the ultimate limits of narrowband and wideband communications.Recently, the wireless standards entity, 3GPP, has introduced new protocols into the standard in order to integrate IoT into the cellular network. A major feature of these protocols is to deploy ultra-low-rate communications in order to address diverse requirements of IoT networks. Such low-rate communication is enabled in the standard by simply allowing a large number of block repetitions, which could be extremely suboptimal from the channel coding perspective. A similar situation arises in wideband scenarios. Wideband channels can essentially provide high data rates, in terms of bits per second. However, individual coded bits experience extremely low signal-to-noise ratios, due to limited signal power and larger attenuations in higher frequencies. In light of the emerging applications in these two extremes of wireless communications, the investigators aim at a comprehensive investigation of channel coding in the low-capacity regime. The specific objectives in this project are as follows: (1) Provide a precise framework for channel coding at low capacity, and characterize fundamental non-asymptotic laws for channel coding in this regime; (2) Develop techniques for the non-asymptotic analysis of wireless wideband and millimeter wave systems; (3) Investigate state-of-the-art codes, including their construction, performance, and efficient decoding, for adaptation to low-rate regimes; (4) Construct efficient concatenated coding schemes including coded repetition and concatenation with low-rate algebraic codes, given diverse requirements of low-capacity applications. Consequently, this project develops a formal connection between information theory, coding theory, and wireless communications resulting in a mathematically precise interface between these areas to address the challenges of narrowband and wideband wireless systems.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

无线网络已通过连接全球数十亿用户来成为我们社会不可或缺的一部分。下一代无线网络（称为5G）有望将数据速率更高的订单传递给数量不断增加的设备。可以将多少网络推向这些进步的基本限制取决于可用的无线频段的数量。为此，两个极端的无线通信变得越来越重要，即窄带和宽带。窄带通信使用少量频率频段可以使数千个设备的连通性，例如图像互联网（IoT）网络中的日常对象，到一个单元塔。使用大量频率的宽带通信在极高的频率下是为5G移动用户提供超高数据速率的关键促进技术。在这两种情况下，各个信息的单个信息都被分配了极低的功率，以使沟通能力产生容量较低的位无线通道。这引入了有关该项目的重点的沟通方面的一系列新的基本问题。该项目的结果将有助于信息和编码理论的基础，以及可以实现窄带和宽带通信最终限制的无线网络的开发。实际上，无线标准实体3GPP将新协议引入了标准中，以便将IoT集成到蜂窝网络中。这些协议的主要功能是部署超低率通信，以解决物联网网络的不同要求。通过简单地允许大量的块重复来实现这种低速率通信，从通道编码的角度来看，这可能是极端最佳的。在宽带场景中出现了类似的情况。宽带通道基本上可以提供高数据速率，而每秒位。但是，由于信号功率有限和较高的频率衰减，单个编码位的信噪比极低。鉴于这两个极端的无线通信中的新兴应用，研究人员旨在全面研究低容量制度中的渠道编码。该项目中的具体目标如下：（1）为低容量的渠道编码提供了一个精确的框架，并表征了基本的非反应定律，用于该制度中的渠道编码； （2）开发用于无线宽带和毫米波系统非反应分析的技术； （3）调查最先进的代码，包括其构造，性能和有效的解码，以适应低利率政权； （4）构建有效的串联编码方案，包括对低容量应用的不同要求，包括编码重复和与低率代数代码的串联。因此，该项目在信息理论，编码理论和无线通信之间建立了正式的联系，从而在这些领域之间在数学上具有精确的界面，以应对窄带和宽带无线系统的挑战。这项奖项反映了NSF的法定任务，并通过使用该基金会的知识分子功能和广泛的影响来评估NSF的法定任务。

项目成果

Low-Complexity Decoding of a Class of Reed-Muller Subcodes for Low-Capacity Channels

低容量信道的一类 Reed-Muller 子码的低复杂度解码

• DOI: 10.1109/icc45855.2022.9838861
• 发表时间: 2022
• 期刊: ICC 2022 - IEEE International Conference on Communications
• 作者: Jamali, Mohammad Vahid;Fereydounian, Mohammad;Mahdavifar, Hessam;Hassani, Hamed
• 通讯作者: Hassani, Hamed

Linear Stochastic Bandits over a Bit-Constrained Channel

• DOI: 10.48550/arxiv.2203.01198
• 发表时间: 2022-03
• 期刊: ArXiv
• 作者: A. Mitra;Hamed Hassani;George Pappas

Channel Coding at Low Capacity

低容量信道编码

• DOI: 10.1109/jsait.2023.3305874
• 发表时间: 2023
• 期刊: IEEE Journal on Selected Areas in Information Theory
• 作者: Fereydounian, Mohammad;Hassani, Hamed;Jamali, Mohammad Vahid;Mahdavifar, Hessam
• 通讯作者: Mahdavifar, Hessam

Fundamental Limits of Two-layer Autoencoders, and Achieving Them with Gradient Methods

两层自动编码器的基本限制以及用梯度方法实现它们

• 发表时间: 2023
• 期刊: Proceedings of the 40th International Conference on Machine Learning
• 作者: Shevchenko, Aleksandr;Kogler, Kevin;Hassani, Hamed;Mondelli, Marco
• 通讯作者: Mondelli, Marco

Federated Neural Compression Under Heterogeneous Data

• DOI: 10.1109/isit54713.2023.10206457
• 发表时间: 2023-05
• 期刊: 2023 IEEE International Symposium on Information Theory (ISIT)
• 作者: E. Lei;Hamed Hassani;S. S. Bidokhti-S.