Collaborative Research: NSF-AoF: CIF: AF: Small: Energy-Efficient THz Communications Across Massive Dimensions

由于目前用于蜂窝通信的频段可用频谱有限，Beyond-5G无线系统有望利用太赫兹频段(0.3-3太赫兹)的大量可用带宽。这将需要在发射机和接收机使用非常大的天线阵列，以补偿在这些高频下信号的强衰减和缺乏穿透。然而，扩展现有的射频技术以在大带宽上运行变得极其复杂、昂贵，并且需要高功耗。因此，需要对射频架构进行彻底的简化；例如，当在巨大的频率和天线尺寸上运行时，不可避免地会牺牲数据转换器的分辨率。这一合作项目解决了与使用超大型阵列相关的理论和实践挑战，这些阵列的每个天线上都连接着数字接收器，并且使用低分辨率的模数/数模转换器。它的目标是实现这项技术的潜力，使太赫兹通信的物理层发生革命性的变化。该项目采用全面的方法，包括分析研究、信号处理方法和网络协议，并利用来自优化、机器学习和贝叶斯推理的严格工具。此外，它将促进参与的美国和芬兰机构之间的研究合作，并支持不同的博士、硕士和本科生在Beyond-5G/6G系统关键技术方面的跨学科发展。该项目将解决Beyond-5G系统中与节能THz通信相关的基本物理层挑战，并被组织为四个相互关联的推动力：i)在链路两端具有低分辨率ADC/DAC的双大规模多输入多输出(MIMO)系统，重点是信道估计、性能分析、(符号级)预编码和解码设计，和硬件非线性；Ii)现实阵列和射频模型下的空间Sigma-Delta处理，结合了二维空间采样，并仔细描述了相互天线耦合和带外发射；iii)初始接入协议，以促进全数字体系结构的实际实现，重点关注同步信号设计和信噪比调谐方案；以及iv)变分贝叶斯推理应用于低分辨率ADC的信道估计和数据检测，在快速变化的环境中取代传统的机器学习模型。通过展示将工作频率推向太赫兹领域的潜力和可行性，该项目将刺激跨学科的研究努力，鼓励低复杂性和可控天线架构的技术进步，并实现高数据速率和低延迟的新无线应用。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。

Due to the limited spectrum available in the frequency bands currently used for cellular communications, beyond-5G wireless systems are expected to exploit the large amount of bandwidth available in the THz band (0.3-3 THz). This will require the use of very large arrays of antennas at both the transmitter and the receiver to compensate for the strong attenuation and lack of penetration of signals at these high frequencies. However, scaling up existing radio-frequency technologies to operate over large bandwidths becomes exceedingly complex, expensive, and demands high power consumption. Thus, radical simplifications in the radio-frequency architecture are needed; for example, a sacrifice in the resolution of the data converters will be inevitable when operating across massive frequency and antenna dimensions. This collaborative project tackles the theoretical and practical challenges associated with using very large arrays with digital receivers attached to each antenna and that employ low-resolution analog-to-digital/digital-to-analog converters. It aims to realize this technology's potential for revolutionizing the physical layer in THz communications. The project adopts a holistic approach that encompasses analytical studies, signal processing methods, and network protocols and leverages rigorous tools from optimization, machine learning, and Bayesian inference. Furthermore, it will foster the research collaboration between the participating US and Finnish institutions and support the cross-disciplinary development of a diverse cohort of Ph.D., Master's, and undergraduate students in key technologies for beyond-5G/6G systems.This project will address the fundamental physical-layer challenges associated with energy-efficient THz communications in beyond-5G systems and is organized into four interconnected thrusts: i) Doubly massive multi-input multi-output (MIMO) systems with low-resolution ADCs/DACs at both ends of the link, focusing on the channel estimation, performance analysis, (symbol-level) precoding and decoding design, and hardware non-linearities; ii) Spatial Sigma-Delta processing under realistic array and radio-frequency models, incorporating two-dimensional spatial sampling and carefully characterizing the mutual antenna coupling and out-of-band emissions; iii) Initial access protocols to facilitate the practical implementation of fully digital architectures, focusing on synchronization signal design and signal-to-noise ratio tuning schemes; and iv) Variational Bayesian inference applied to channel estimation and data detection with low-resolution ADCs, to be used in lieu of traditional machine learning models in rapidly time-varying environments. By demonstrating the potential and feasibility of pushing the operating frequencies to the THz realm, the project will stimulate cross-disciplinary research efforts, encourage technological advancements in low-complexity and controllable antenna architectures, and enable new wireless applications with high data rates and low latencies.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.