CRII: NeTS: Next Generation Spectrum Measurement Algorithms and Infrastructures

当前的独占频谱分配模式正在造成频谱资源的人为稀缺，这对网络性能和用户体验产生了巨大影响。未充分利用的频段为更高效的共享频谱接入提供了机会，最近将政策制定者、行业领导者和学术研究人员聚集在一起，为下一代频谱管理制定了议程。这种进步的一个关键推动因素是对频谱使用的深刻理解。该奖项将支持寻求影响科学界、政策领域和公众的研究，为有效的频谱测量和表征提供方法和系统设计，并为未来移动无线网络和频谱政策的设计提供信息。这项研究的成果将包括在PI的本科和研究生课程中，展示计算机科学的积极政策和社会影响，并向不同的学生群体推广该学科。该项目的目标是利用信号处理和传播、机器学习和大规模测量的基础知识来设计快速和健壮的算法，并为频谱测量基础设施设计提供信息。主要的研究兴趣将是算法和基础设施之间的紧密相互依赖。该项目将通过两项研究任务实现这一目标：(I)设计和开发用于自适应频谱感知、频谱数据管理和频谱表征的频谱测量算法；(Ii)评估频谱测量基础设施，阐明频谱传感器属性对频谱表征质量的影响，并为设计传感器成本/移动性混合测量基础设施提供参考。这项研究将在算法设计和基础设施之间架起一座桥梁。它将推出首个用于频谱测量和表征的整体框架，以促进下一代频谱管理。自适应的、数据驱动的算法的设计将促进快速有效的光谱测量和表征。探索频谱感知基础设施中的成本和移动性权衡将有助于高效的端到端基础设施设计，从而最大限度地降低成本，同时最大化频谱感知的学习结果。

The current paradigm of exclusive spectrum allocation is creating artificial scarcity of spectrum resources that has a dramatic impact on network performance and user experience. Underutilized bands provide an opportunity for more efficient, shared spectrum access that has recently brought together policymakers, industry leaders and academic researchers to set an agenda for next-generation spectrum management. A critical enabler of such advances is deep understanding of spectrum use. This award will support research seeking to impact the scientific community, the policy domain and the public, by providing methods and informing system designs for efficient spectrum measurement and characterization, and informing the design of future mobile wireless networks and spectrum policy. Products from this research will be included in the PI's undergraduate and graduate courses, demonstrating the positive policy and societal impact of computer science and promoting the discipline to diverse groups of students.The goal of this project is to employ fundamental knowledge in signal processing and propagation, machine learning, and large-scale measurement to devise fast and robust algorithms and inform spectrum measurement infrastructure design. Of main research interest, will be the tight-knit mutual dependence between algorithms and infrastructures. The project will achieve this goal in two research tasks: (i) design and development of spectrum measurement algorithms for adaptive spectrum sensing, spectrum data management and spectrum characterization; and (ii) evaluation of spectrum measurement infrastructures that will illuminate the impact of spectrum sensor properties on spectrum characterization quality and will inform the design of mixed sensor cost/mobility measurement infrastructures. This research will bridge the gap between algorithm design and infrastructures. It will introduce the first holistic framework for spectrum measurement and characterization to facilitate next generation spectrum management. The design of adaptive, data-driven algorithms will facilitate rapid and efficient spectrum measurement and characterization. Exploration of the cost and mobility trade-offs in spectrum sensing infrastructure will inform efficient end-to-end infrastructure design that minimizes cost while maximizing the learning outcomes of spectrum sensing.