Collaborative Research: SWIFT: SMALL: Autonomously Reconfigurable Hardware-Reduced Wideband Transceivers for Efficient Passive-Active Spectrum Coexistence

商用和军用无线系统、射电天文观测站、天气雷达系统和其他应用，通常被称为“无源用户”，需要在干扰有限的安静电磁环境中运行。无源用户通常努力观察由远距离非协调发射机发射的微弱信号，并且它们与其他源的频率或时间分离可能是不可能的，因为它们经常连续地利用大范围的无线电频谱来检测和监控在不同频率具有电磁存在的物理过程。为了充分获得5G及以后频谱共存的好处，需要采用在两个关键目标之间取得最佳平衡的被动用户保护方法，即可靠地保护被动用户免受干扰和最大限度地增加主动用户的频谱接入机会。该项目通过开发一种新型的低成本硬件和多参数可重构超宽带收发器来解决频谱共存方面的基本挑战，该收发器可在较宽的频率范围内优化被动-主动频谱共享。这种新型收发器的便携性和适应性使其对下一代移动无线平台具有吸引力，包括各种自主地面和/或机载平台、蜂窝基站、无人机载和卫星通信系统，特别是影响5G、Wi-Fi和互联自治的未来应用。通过这个项目，PIS建议通过专门的外展努力和课程开发，培养针对西班牙裔、女性和其他科学、技术、工程和数学(STEM)中代表性不足群体的本科生、研究生和博士后。该项目汇集了一支跨学科的研究团队，他们具有互补的专业知识，从射频(RF)前端硬件设计，到物理(PHY)和介质访问控制(MAC)层优化，再到人工智能(AI)理论和实践。为了避免模拟和混合波束形成架构带来的硬件限制和缺乏可重构性，该项目开发了一种新型的数字超宽带波束形成架构，该架构能够实现：1)通过物理层和MAC层稳健的空时频感知和跨层优化，实现高效的频谱利用和无干扰的被动-主动共存；2)针对跨5G频段的超宽带操作，自主地实现硬件多参数可调；以及3)通过新型人工神经网络辅助的码复用阵列前端，实际实现低成本、通用的硬件精简的无线系统。稳健的频谱感知涉及新的L1范数主成分分析，以评估收集/感知的频谱数据的质量(有效性和完整性)，并产生高置信度的功率传播和活跃频谱用户的空间坐标地图。该项目利用高质量的无线电地图、自治子阵列、多频段和跨宽带的多参数可重新配置，并结合强化学习策略来解决跨层PHY/MAC和前端控制问题，以实现和谐的被动和主动频谱共存。该项目的全部成果预计将是一个可自主重新配置的硬件缩减平台，它将允许在小尺寸软件定义的无线电平台中集成超宽带侦听和波束形成功能。这种新型的低成本、多功能、硬件减少的无线收发器将多个无线链路集成到一个实用的单个轻量级封装中，通过应用神经网络来消除在多个非正交扩展信号路径的多路复用过程中产生的通道间干扰的线性和非线性分量。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。

Commercial and military wireless systems, radio astronomy observatories, weather radar systems, and other applications, frequently referred to as “passive users,” need to operate in quiet electromagnetic environments with limited interference. Passive users commonly strive to observe faint signals emitted by distant non-coordinating transmitters and their frequency or time separation from other sources may not be possible as they often continuously utilize large swaths of radio spectrum to detect and monitor physical processes that have electromagnetic presence at different frequencies. To fully reap the benefits of spectrum coexistence in 5G and beyond, there is a need to employ passive-user protection approaches that strike an optimal balance between two key objectives, namely reliably protect passive users from interference and maximize spectrum access opportunities for active users. This project addresses foundational challenges in spectrum coexistence by developing a novel low-cost hardware-reduced and multi-parameter reconfigurable ultra-wideband transceiver that optimizes passive-active spectrum sharing across a broad frequency range. The portability and adaptability of this new transceiver makes it attractive for next-generation mobile wireless platforms, including a variety of autonomous ground and/or airborne platforms, cellular base-stations, unmanned airborne and satellite communication systems and specifically impact 5G, Wi-Fi and future applications of connected autonomy. Through this project the PIs propose to train undergraduate, graduate, and postdoctoral students targeting Hispanic, women, and other underrepresented groups in science, technology, engineering and mathematics (STEM) through specialized outreach efforts and curriculum development.The project brings together an interdisciplinary team of researchers with complementary expertise ranging from Radio Frequency (RF) front-end hardware design, to Physical (PHY) and Medium-Access Control (MAC) layer optimization, to artificial intelligence (AI) theory and practice. To avoid hardware constraints and lack of reconfigurability imposed by analog and hybrid beamforming architectures the project develops a novel digital ultra-wideband beamforming architecture that enables: 1) Efficient spectrum utilization and interference-free passive-active coexistence through robust space-time-frequency sensing and cross-layer optimization at the PHY and MAC layers; 2) autonomous hardware multi-parameter tunability for ultra-wideband operation across 5G bands; and 3) practical realization of low-cost, versatile hardware-reduced wireless systems through new artificial-neural-network-aided code multiplexed array front-ends. Robust spectrum sensing involves new L1-norm principal-component analysis to assess the quality (validity and completeness) of the collected/sensed spectrum data and produce high-confidence power propagation and spatial coordination maps of active spectrum users. The project leverages high-quality radio maps, autonomous sub-arraying, multi-band, and multi-parameter reconfigurability across large bandwidths and incorporates reinforcement learning strategies to address the cross-layer PHY/MAC and front-end control problem for harmonious passive-active spectrum coexistence. The complete outcome of this project is expected to be an autonomously reconfigurable hardware-reduced platform that will allow integration of ultra-wideband sensing and beamforming functionalities in small-form-factor software-defined radio platforms. This new class of low-cost, versatile, hardware-reduced wireless transceivers that integrate multiple radio chains into a practical single lightweight package is enabled by the application of neural networks to cancel both linear and non-linear components of inter-channel interference that arise during the multiplexing of multiple non-orthogonal spread signal paths.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.