Collaborative Research: SWIFT: SMALL: Learning-Efficient Spectrum Access for No-Sensing Devices in Shared Spectrum

合作研究：SWIFT：SMALL：共享频谱中无感知设备的学习高效频谱访问

• 批准号: 2030026
• 负责人: Jing Yang
• 金额: $ 22万
• 依托单位: Pennsylvania State Univ University Park
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-15 至 2024-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2030026&HistoricalAwards=false
• 关键词: Collaborative; Research; SWIFT; SMALL; Learning

This project develops a novel online learning based framework for distributed low-cost devices to efficiently and effectively access the shared spectrum without spectrum sensing. It specifically focuses on no-sensing devices that do not have the powerful radio-frequency (RF) components to enable wideband spectrum sensing, and addresses the cross-technology spectrum access problem in a decentralized setting. A pertinent application the proposed solution addresses is the dynamic spectrum access of Internet-of-Things (IoT) devices that are deployed in either unlicensed or lightly licensed spectrum, in which the distributed IoT devices need to coexist with other active systems. The no-sensing spectrum access and sharing framework has the potential to revolutionize the operation and management of modern and future wireless networks, considerably enhance the spectrum utilization efficiency, and dramatically alleviate the constantly increasing pressure on the limited radio spectrum. The cross disciplinary nature of the research would naturally translate into case studies and projects in a number of undergraduate and graduate level courses taught by the PIs in areas of communications, machine learning, and networking.This project aims to develop a suite of online learning based spectrum access algorithms for no-sensing devices to coexist with other active systems. The first study focuses on improving the learning efficiency by introducing the best arm identification framework and proposing meta-learning and good channel identification algorithms. The second thrust is devoted to designing spectrum access mechanisms that can seamlessly integrate hybrid automatic repeat request (HARQ). Novel algorithms will be designed to learn the optimal sequence of channels for possible retransmissions, and enhanced for fine-grained control that captures the coding level behavior of HARQ. The last thread of investigation considers multi-user multi-technology coexistence and will develop implicit-communication based distributed spectrum access algorithms. Finally, a thorough validation of the algorithms and spectrum access schemes will be performed using a lab testbed and real-world datasets.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.

该项目为分布式低成本设备开发了一种新的基于在线学习的框架，以在没有频谱感知的情况下高效地访问共享频谱。它特别关注那些没有强大的射频(RF)组件来实现宽带频谱感知的无感知设备，并在分散设置下解决了跨技术的频谱接入问题。推荐的解决方案解决的一个相关应用是部署在未经许可或轻度许可频谱中的物联网(IoT)设备的动态频谱访问，在这种情况下，分布式物联网设备需要与其他活动系统共存。无感知频谱接入和共享框架有可能彻底改变现代和未来无线网络的运营和管理，显著提高频谱利用效率，并极大地缓解有限无线电频谱不断增加的压力。这项研究的跨学科性质自然会转化为一些本科生和研究生水平的课程中的案例研究和项目，这些课程由私人投资机构教授，涉及通信、机器学习和网络领域。该项目旨在开发一套基于在线学习的频谱接入算法，用于无感知设备与其他主动系统共存。第一个研究重点是通过引入最佳ARM识别框架和提出元学习和良好的信道识别算法来提高学习效率。第二个推力致力于设计能够无缝集成混合自动重传请求(HARQ)的频谱接入机制。新的算法将被设计来学习用于可能的重传的最优信道序列，并针对捕获HARQ的编码级行为的细粒度控制进行增强。最后一条研究思路是考虑多用户多技术共存，开发基于隐式通信的分布式频谱接入算法。最后，将使用实验室试验台和真实世界数据集对算法和频谱接入方案进行彻底验证。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Cascading Bandits with Two-Level Feedback

具有两级反馈的级联 Bandits

• DOI: 10.1109/isit50566.2022.9834892
• 发表时间: 2022
• 期刊: 2022 IEEE International Symposium on Information Theory (ISIT
• 作者: Cheng, Duo;Huang, Ruiquan;Shen, Cong;Yang, Jing
• 通讯作者: Yang, Jing

Joint User Association and Wireless Scheduling with Smaller Time-Scale Rate Adaptation

联合用户关联和较小时标速率自适应的无线调度

• 发表时间: 2023
• 期刊: and Wireless Networks (WiOpt
• 作者: Wu, X.;Yang, J.;Zeng, H.;Li, B.
• 通讯作者: Li, B.

Near-optimal Conservative Exploration in Reinforcement Learning under Episode-wise Constraints

• DOI: 10.48550/arxiv.2306.06265
• 发表时间: 2023-06
• 期刊: Mathematics
• 影响因子: 2.4
• 作者: Donghao Li;Ruiquan Huang;Cong Shen;Jing Yang

Heterogeneous Multi-player Multi-armed Bandits: Closing the Gap and Generalization

• 发表时间: 2021-10
• 作者: Chengshuai Shi;Wei Xiong;Cong Shen;Jing Yang

Provably Efficient Offline Reinforcement Learning with Perturbed Data Sources

• DOI: 10.48550/arxiv.2306.08364
• 发表时间: 2023-06
• 期刊: ArXiv
• 作者: Chengshuai Shi;Wei Xiong;Cong Shen;Jing Yang