Collaborative Research: CCSS: Coding for 5G and Beyond: Limits and Efficient Algorithms

合作研究：CCSS：5G 及以后的编码：限制和高效算法

• 批准号: 1711056
• 负责人: Joerg Kliewer
• 金额: $ 19.24万
• 依托单位: New Jersey Institute of Technology
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-09-01 至 2022-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1711056&HistoricalAwards=false
• 关键词: Collaborative; Research; CCSS; Coding; 5G

During the next couple of years, an exponential increase of data use per capita will be experienced. For example, predictions by the cellular industry state that the global per-capita information usage will grow from 15 GB in 2014 to around 37 GB in 2019. Without significant technological advances to increase its capacity, the existing telecommunications infrastructure will be unable to support this vast data increase. The advent of modern error-correcting codes, such as turbo codes, low-density parity-check codes, and polar codes, has represented a quantum leap in error-correcting performance for wireless systems, allowing reliable communication of information over noisy channels at data rates close to capacity. However, improvements in this direction have been mostly limited to the point-to-point case, and maximizing gains for the point-to-point channel setup will not be sufficient to satisfy the high throughput and low delay requirements of emerging communication systems, in particular of 5G cellular systems and beyond. This project aims to tackle these issues by extending coding schemes to the wireless network setting, thereby yielding additional throughput gains beyond the classical point-to-point case. The proposed research promises to provide a significant transformative impact on many critical applications employing reliable networked wireless communication, for example in the fields of healthcare, environmental monitoring, finance, and so on.In this project, a comprehensive framework of practical low-complexity, low-latency, capacity-approaching sparse graph codes is proposed. These codes are able to leverage network gains for error correction, thereby significantly reducing the amount of transmitted data. The project aims to study the fundamental limits of these schemes as well as to investigate practical coding algorithms to approach these limits. The proposed research involves several fundamental themes related to the application of sparse graph-based and polar codes to emerging future communication systems which are not present in previous studies: analysis and design of nested codes for canonical network communication problems; analysis of how performance scales with various code and decoder design parameters; a theoretical understanding of the implementation complexity versus performance trade-offs between algebraic and random codes; an analytical investigation of iterative decoding failure events; and the development of a novel high-speed field-programmable gate array hardware decoding architecture.

在接下来的几年里，人均数据使用量将呈指数级增长。例如，蜂窝行业的预测称，全球人均信息使用量将从2014年的15 GB增长到2019年的约37 GB。如果没有重大的技术进步来增加其容量，现有的电信基础设施将无法支持这种巨大的数据增长。诸如Turbo码、低密度奇偶校验码和极性码等现代纠错码的出现代表了无线系统纠错性能的巨大飞跃，允许在噪声信道上以接近容量的数据速率进行可靠的信息通信。然而，在这一方向上的改进主要限于点对点情况，最大限度地提高点对点信道设置的增益将不足以满足新兴通信系统的高吞吐量和低延迟要求，特别是5G蜂窝系统和更高的系统。该项目旨在通过将编码方案扩展到无线网络设置来解决这些问题，从而在经典的点对点情况下产生额外的吞吐量收益。提出了一种实用的低复杂度、低延迟、逼近容量的稀疏图形码框架，该研究有望对许多采用可靠的网络无线通信的关键应用，如医疗、环境监测、金融等领域产生重大的变革影响。这些代码能够利用网络增益进行纠错，从而显著减少传输的数据量。该项目旨在研究这些方案的基本限制，并研究实用的编码算法以接近这些限制。这项研究涉及几个与稀疏图和极性码在未来新兴通信系统中的应用相关的基本主题，这些主题在以前的研究中没有出现：针对规范网络通信问题的嵌套码的分析和设计；分析性能如何随不同的码和解码器设计参数的变化而变化；从理论上理解代数码和随机码的实现复杂性与性能的权衡；对迭代译码失败事件的分析研究；以及开发一种新型的高速现场可编程门阵列硬件译码架构。

项目成果

Learned Scheduling of LDPC Decoders Based on Multi-armed Bandits

基于多臂老虎机的LDPC解码器的学习调度

• 发表时间: 2020
• 期刊: IEEE International Symposium on Information Theory
• 作者: Habib, Salman;Beemer, Allison;Kliewer, Joerg
• 通讯作者: Kliewer, Joerg

Algebraic Optimization of Binary Spatially Coupled Measurement Matrices for Interval Passing

• DOI: 10.1109/itw.2018.8613339
• 发表时间: 2018-09
• 期刊: 2018 IEEE Information Theory Workshop (ITW)
• 作者: Salman Habib;J. Kliewer

A generalized algebraic approach to optimizing SC-LDPC codes

• DOI: 10.1109/allerton.2017.8262802
• 发表时间: 2017-07
• 期刊: 2017 55th Annual Allerton Conference on Communication, Control, and Computing (Allerton)
• 作者: Allison Beemer;Salman Habib;C. Kelley;J. Kliewer

Learning to Decode: Reinforcement Learning for Decoding of Sparse Graph-Based Channel Codes

学习解码：基于稀疏图的信道码解码的强化学习

• 发表时间: 2020
• 期刊: Advances in neural information processing systems
• 作者: Habib, Salman;Beemer, Allison;Kliewer, Joerg
• 通讯作者: Kliewer, Joerg

Belief Propagation Decoding of Short Graph-Based Channel Codes via Reinforcement Learning

通过强化学习对基于图的短信道码进行置信​​传播解码

• DOI: 10.1109/jsait.2021.3073834
• 发表时间: 2021
• 期刊: IEEE Journal on Selected Areas in Information Theory
• 作者: Habib, Salman;Beemer, Allison;Kliewer, Jorg
• 通讯作者: Kliewer, Jorg