NSF Award
1634300
Radio frequency spectrum is finite and yet inefficiently used -- resulting from prior policy regimes that dispensed spectrum via methods initially such as administrative allocation and later via market mechanisms such as auctions. Accordingly, a key imperative today is to enable Spectrum Sharing whereby new services are allowed to operate in bands hitherto allocated to licensed incumbents (primary users) under new sharing rules, driven by a Presidential mandate directing 500 MHz currently held by various Federal networks to be re-purposed by 2020 for civilian use. A key enabler of any engineering solution for the above is deployment of (low cost) hardware in conjunction with appropriate networking and storage resources and a software infrastructure that allows for persistent monitoring of spectrum usage. This pilot project will produce the first calibrated large-scale data sets (over frequency, time and region) of spectrum usage archived for post-processing. Such a database will be a unique national resource for any future spectrum engineering efforts; its impact will extend well beyond the academic community. The data-driven research that it enables will provide valuable inputs for spectrum policymakers charged with formulation of new spectrum sharing rules.<br/><br/>This project significantly extends the state-of-the art in spectrum monitoring: first and foremost, it will result in the first wide-area, persistent and taskable spectrum monitoring and data collection infrastructure without using expensive high-end sensors, but still capable of gathering unprecedented amount of good quality spectrum data that will be stored and made publicly available for further investigations by research community. Further, the design will constitute the first step towards future networked monitoring infrastructure that can allow for distributed experimentation by researchers who will have the ability to task desired sensor sub-sets simultaneously for data acquisition in desired spectrum bands. This will enable data mining of spectrum usage patterns as a function of both space and time, which is currently not possible.
